Proceedings of the NASA, Industry, University, General Aviation Drag Reduction Workshop

Overview of Configuration Drag

Part I - Cost Considerations for Aircraft Configuration

Changes

R. Turnlinson, Beech Aircraft Corporation . I ~ . , . . 331 '

Part I1 - Aerodynamic Considerations

J . Roskam, University of Kansas . . . . . . . . 337 6

0.2 Learjet Model 25 Drag Analysis

R. Ross and R. D. Neal , Gates Learjet Corporation . . . 353

8.3 Problems in Propulsion System Integration

W. Henderson and J . Runckel , NASA Langley Research

Center . . . . . . . . . . . . . . . . . . 365

8.4 Propulsion/Airframe Integration

0 . Mikkelson , NASA Lewis Research Center . . . . . . 387

8.5 Determination of the Level Flight Performance of Propeller-Driven Aircraft

E. J. Cross, Jr. , Mississippi State University . , . . . . 403

9. ADDITIONAL PAPERS RECEIVED AFTER THE CONFERENCE . . . 407

Possible Applications of Soaring Technology to Drag 9.1 Reduction i n Powered General Aviation Aircraft

J. H. McMasters and G. M. Palmer, Purdue University . . 409

9.2 Minimum Vertical Tail Drag

E. E. Larrabee, Massachusetts Institute of Technology . . 2

...

I l l

Preceding Page Blank

his document contains all papers presented at the "NASA/Industry/ University General Aviation Drag Reduction Workshop," held at the University of Kansas July 14-16, 1975. The conference was sponsored by NASA Langley Research Center under NASA Grant NSG 1175. The sequence of papers in this document i s the same as that found on the Program Outline, Chapter 2, The actual papers are distributed over Chapters 3 through 8. A number of papers were received after the conference. Several of these were judged to be of sufficient interest to the subject of general aviation drag reduction to include them in this document i n Chapter 9. In addition, a summary of all technical discussions which followed the papers i s found i n Chapter 10.

The purpose of the conference was to: (1) Identify the state-of-the art, and (2) Formulate a NASA/University R & D program aimed a t achieving significant drag reductions i n the near term future.

Chapter 11 contains a summary of recommendations for drag reduction research aimed at improving general aviation airplanes.

Preceding Page Blank

2. P R O G

Preceding Page Blank

2, F I 0 G .Monday, July 14, 1975

8:15 - 8:45 Registration, Lobby, Nichols Hall

8:45 - 9:OO Welcome and Introduction

J. Roskam, university of Kansas

9:OO - 9:45 General Overview of Drag

S. B. Anderson, NASA-Ames Research Center

&Session I - Status of Draa Prediction Methods

9:45 - l0:30 Overview of Drag Prediction Methods

D. Ruhmel, Cessna Aircraft Company

10:30 - 10:45 Break

10:45 - 11:30 Brief Presentations by Industry, Universities and NASA

The following presentations are scheduled: 1. Prospects and Time Tables for .Analytical Estimation of the Drag of Complete Aircraft Configurations F. 0. Srnetana, North Carolina State University Summary of Drag Cleanup Tests i n the NASA Langley 2.

Fu II -Sca I e Tunnel M. 0. McKinney, NASA Langley Research Center Simp1ified Theoretical Methods for Aerodynamic Design 3.

J. Tulinius, NASA Langley Research Center Open Discussions. Formulation of research and develop-

11:30 - 12:OO

ment work needed i n the area of drag prediction methods.

12:oo - 2:oo Dutch Treat Luncheon - Centennial Room, Student Union

Bu i 1 ding Speaker: 0 . W. Nicks, NASA Langley Research Center Topic: Drag Reduction/Back to Basics

+Session I1 - Fuselage Drag

2:OO - 2 : 3 0 Overview of Fuselage Drag

J . Roskam, University of Kansas

2:30 - 3:OO Brief Presentations by Industry, Universities and NASA

The'fol lowing presentations are scheduled: 1. Propeller Blockage Research Needs R. Tumlinson, Beech Aircraft Corporation Preservation of Wing Leading Edge Suction a t the 2.

Plane of Symmetry as a Factor i n Wing-Fuselage Design E. E. Larrabee, Massachusetts Institute of Technology

Preceding Page Blank

inia Polytechnic Institute Break

3:15 - 4:30 Open Discussions. Formulation of research and development

work needed in the area of fuselage drag.

4:30 - 5 3 0 Visit to KU Flight Research Laboratory Facilities

4:30 Flight Simulator Lab, Nichols Hall 5:OO Wind Tunnel & Tornado Lab, Learned Hall

6:OO - 7:OO Dutch Treat Sociai Hour, Ramada Inn

7: 00 Dutch Treat Dinner, Ramada Inn Speaker: R. D. Neal, Gates Learjet Corporation Topic: The Economic Impact o f Drag i n General Aviation aTuesday, July 15, 1975

#Session 111 - Wing Drag

Methods for Reducing Wing Drag and Wing-Nacelle Inter- 8:30 - 9:OO

ference T. C. Kelly, NASA Langley Research Center

Brief Presentations by Industry, Universities and NASA 9:OO . . 10:3O

Scheduled are the following brief presentations: 1. Drag Reduction through Higher Wing Loading D. L. Kohlman, University of Kansas 2. Use of a Pitot Static Probe for Determining Wing Section Drag i n Flight L. C . Montoya, P. S. Bikle and E . Saltzman, NASA Flight Research Center Flight Test Results with an Ogee Wing Tip 3.

J . Vogel, Beech Aircraft Corporation

Wing-Tip Vanes as Vortex Attenuation and Induced 4.

Drag Reduction Devices W. H. Wentz, Jr., Wichita State University 5. Wing Tip Vortex Drag V. U. Muirhead, University of Kansas

10:30 - 10:45 Break

10:45 - 12:OO Open Discussions. Formulation of research and development

work needed in the area of wing drag.

12:OO - 2:OO Dukh Treat Luncheon - Centennial Room, Student Union

Speaker: R. Winblade, NASA Headquarters Topic: NASA Light Aircraft Performance Research: Opportunities and Limitations {Cancelled; S. Anderson showedand narrated a V/STOL movie instead.)

2:45 - 345 Brief Presentations by Industry, Universities and NASA

1 . Installation Drag Considerations Other than External Nacelle and Interference Drag as Related to Turboprop and Turbofan Engines G . Burnett, Garrett AiResearch Manufacturing Co.

of Arizona 2. Nacelle Drag Reduction: An Analytical Guided Exper i menta I Program F. Smetana, North Carolina State University Cooling Drag Associated with General Aviation Pro- 3.

pulsive Systems

E . J. Cross, Jr . , Mississippi State University

Propellers of Minimum Induced Loss, and Water Tunnel 4.

Tesh of Such a Propeller E . E . Larrabee, Massachusetts Institute of Technology

3:45 - 4 : O O Break

4:OO - 500 Open Discussions. Formulation of research and development

work needed in the area of nacel I e and interference drag.

Dutch Treat Social Hour

6:OO - 7:OO

Dutch Treat Dinner, Ramada Inn 7:OO .Wednesday, July 16, 1975

jcSession V - Trim Drag

8:30 - 9: 15 Overview o f Tr.im k a g

J . Roskam, University of Kansas

Brief Presentations by Industry, Universities and NASA

9:15 - 10:15

Scheduled are the following brief presentations: 1. Trim Drag Research ResuIfs H. Chevalier, Texas A & M University 2. Reduction of Trim Drag in General Aviation Airplanes F. H. Lutz, Virginia Polytechnic Institute 3. Trim Drag i n the l i g h t of Munk's Stagger Theorem E . E . Larrabee, Massachusetts Insitute of Technology

10:15 - 10:30 Break

10;30 - 11 :30 Open Discussions. Formulation o f research and development

work needed in the area of trim drag.

11:30 - 1:OO Dutch Treat Luncheon - Centennial ROOM, Student Union

No Speaker Scheduled

art I - Cost Considerations; R. Tumlinson, Corporation

Part I1 - Aerodynamic Considerations; J . Roskam, University

of Kansas

Brief Presentations by Industry, Universities and NASA 1 :30 - 2:30

Scheduled are the following brief presentations: Learjet Model 25 b a g Analysis 1.

R. Ross and R. Neal, Gates Learjet Corporation 2. Problems i n Propulsion System Integration W. Henderson and J. Runckel, NASA Langley Research Center 3. Propu!sion/Airfrarne Integration D. Mikkelson, NASA Lewis Research Center Determination of the Level Flight Performance of Pro- 4.

peller -Driven Aircraft E . J. Cross, Jt., Mississippi State University

2:30 - 3:15 Open Discussions. Formulation of research and developmen)

work needed in the area o f complete configuration drag.

3:15 - 3:30 Closing Remarks

J . Stickle, NASA Research Center and J. Roskam, University of Kansas The importance of reducing drag for general aviation aircraft is increasingly evident for the reasons noted in Figure 1 . This includes rising fuel costs and the demand for improved performance to meet fore ign cornpet it ion. Equal I y important is the impact of more stringent noise and pollution standards because these factors indirectly affect aerodynamic performance. Although the general principles of how t o achieve drag reduction are known to aircraft designers, applications to general aviation aircraft are a significant challenge because this aircraft category is particularly sensitive to costs, maintenance, marketing, safety utility, and even stability and control.

How much do we really know about potential drag reductions for a typical high-performance business aircraft? A casual inspection of a current twin shows a n abundance of brazier head rivets on all parts of the aircraft, several large external antennas, a lack of wing-fuselage filleting, lapped skin joints, many air inlets at obviously undesirable aerodynamic locations, and a single large-diameter exhaust pipe protruding at close to 9 0 ' to the airstream. O n one twin turboprop aircraft, NACA flush inlets were located on each engine nacelle, some seven separate obviously of quesfionable value for pressure recovery. Although it is recognized that little systematic research on drag for current aircraft configurations has been.

conducted recently, many of the results of early NACA research c a n be usefully applied t o current aircraft. Obviously, there is little similarity between the blunt, radial-engine transport aircraft of the late 1930's, for which most of the early research was conducted, to today's sleek business jet, s o f e w would question the need for additional research.

As noted in the program for this workshop, it i s timely t o identify the state-of- the-art o n aerodynamic drag reduction and develop a program plan for achieving meaningful results. There are, of course, many elements making up the total drag of an aircraft, including fuselage, wing, nacelles, >trim, interference, tail, and cooling drag. The various topics to be covered in the next three days are shown on F igure 2.

Note that although cooling drag can be a large percentage of total drag (as high as 25%), it has previously been covered in a NASb&nivenity/Industry workshop and will not be considered explicitly at this workshop. As noted in Figure 3, the purpose of this paper is to review the relative drag contributions of these various elements, nsiderat~o~ by speakers who w i l l Basic Sources of Drag It is important to identify the basic sources of drag in order to gain a better understanding of how improvements in performance can be made. Shown on Figure 4 are the following: (1) skin friction due to the air molecules rubbing the surface, the magnitude being a function of the flow conditions (laminar or turbulent) and the amount of wetted area; (2) induced flow or vortex flow primarily a function of wing aspect ratio; and {3) pressure effects associated with the profile o r form of various parts of the aircraft.

Shown on Figure 5 i s the variation of flat plate drag coefficient based on wetted area with Reynolds number for fully turbulent and laminar flow conditions.

Note that at large Re numbers typical of flight cruise conditions, the drag associated w i i h turbulent flow i s ten times higher than for laminar flow. In another example of the effect of flow conditions, Figure 6.compres the equivalent drag of a laminar flow airfoil and a circular wire. If nothing else, this i s an incentive to avoid using exposed landing wires.

Drag Prediction Techniques Moving along to the first topic of our workshop, the various drag prediction techniques in use toduy are noted in Figure 7. The empirical approach takes udvmtage of semi-analytical methods in which wind-tunnel and flight-test results of similar type aircraft are factored in to establish a data base. Wind tunnel measurements of drag for a new design are usually made, particularly for high-performance aircraft.

Extrapolation of smafl-scale (low Re n o . ) data to flight conditions can be difficult when including power effects and the accuracy of how well the small-scale model represents the actual aircraft. Finally, theoretical estimates, although used extensively in the past, h w e become more popular because of the availability of large capacity digital computers. Solutions of 3-dimensional viscous flow effects appear to remain a challenge even with very large (and expensive) digital computers such as the ILLIAC I V based at the NASA Ames Research Center.

n example of results from drag prediction methods developed ut ~ASA-Ames dynam ics subrout ine ca culates a series of factors which are used to establish drag values. Form factors are used for each component to represent drag increases above that of a flat plate to account for 3-0 effects, interference, roughness, and excrescences. These calculations were made for the

Learjet , Citation, Cessna 340, Piper Arrow, and Cessna 1.50. Note first , not

unexpectedly, that the wing and Fuselage are responsible for the largesf source of drag.

Of interest in the last column i s the amount to be added to match flight values of drag. This item varies greatly, gbing from less than 2 percent for the Learjet to 37 percent for the Cessna 150. improvements are needed to more accurately account for such factors as 3-0 effects, cooling drag, landing gear, slipstream drag, etc.

Factors Influencing Fuselage Drag I n the next item of our workshop agenda, Figure 9 gives several factors which affect fuselage drag. The surface conditions are very important because of the large wetted area. Windshield shape can signifikantly affect total drag at the higher Mach numbers. Fuselage shape in terms of fineness ratio, nose shape and rear-end shape Shown in Figure 10 i s the effect o f afterbody contraction must be carefully considered.

ratio on drag. The contraction ratio must be greater than 2.0 to avoid a drag increase.

A similar Consideration must be given in the vertical plane,.

Factors Influencing Wing Drag Figure 11 lists several factors which are considered in selecting a wing for a new aircraft design. A large background of data is available from N A C A research on airfoil sections and newer types such as the GAW-I airfoil to challenge the designer in selecting the correct airfoil section for his aircraft. The NASA has underway a program on airfoil development aimed primarily at optimizing airfoils for specific operating conditions. Thickness rat io effects are generally well-documented. Planform and aspect ratio effects are also important as influenced by structural considerations.

Wing-tip effects on induced drag w i l l be covered specifically in a Langley Research Center paper describing the trade-offs on using "Winglets. I' Another reminder of the importance of surface conditions and thickness ratio on drag i s given in Figure 12. These NACA data tend to exaggerate the effect of roughness because the lower curve represents a mirror-finish surface condition. The ance jets is to use icker airfoil sect ions for e considerat ions.

Factors Influencing Trim Drag O f the various factors shown in Figure 13 which affect trim drag, tail location and static stability have recently been given increased attention. A tail location out of the slipstream ("TI' tail designs) offer some drag decrease, and canard horizontal tail locations have appeared on experimental aircraft. In consideration of the small percentage of the tail surfaces to total drag indicated previously, one must be careful

not t o compromise stability and control in looking for performance imporvements . In

this connection the control configured-vehicle (CCV) and relaxed static stabif ity have received attention recently. An illustration of the effect of reducing static margin the horizontal tail area required i s shown in Figure 14. These curves indicate the on variation of tail size with static margin to trim out the wing-fuselage pitching moment and the tail area needed for maneuvering. To achieve the minimum tail area and therefore the least amount o f drag, the static margin must be slightly aft of the neutral

( d s d k = 0) but ahead of the maneuver point ( d F J d , t i , , = 0). Obviously, some

point L for of stability augmentation must be provided to meet the FAR if minimum tail size i s desired. At this point one would logically question the merits of reducing static margin for moot General Aviation aircraft.

Considerations for Drag of Complete Aircraft In the final analysis, drag of the complete configuration is the most difficult to rationalize. As noted in Figure 15, cost i s a factor that must be considered in each aspect of aerodynamic drag reduction. Cost aspects will be discussed in a paper later in the workshop. In this regard use of composites may offer promise in that extremely smooth surfaces with attendant low drag can be achieved without high-cost manufacturing techniques. The second point, aerodynamic drag of the complete configuration, must take into account items such as wing nacelle and tail location, fuselage camber, wing and nacelle incidence, wing loading, cruise I ift coefficient, etc. This area w i l l be covered also on the last day of the workshop. The next item, propulsion system integration, is an important area, particularly for higher performance aircraft. Nacelle size and location can significantly affect high subsonic Mach number performance, as w i l l be discussed by NASA Lewis Research Center. Fabrication details, the next item, must be considered in the I ight of cost and aircraft appearance. A t only has the potential for higher ~ e r f ~ m ~ ~ c e , is the relative magnitude of the ~ m p ~ r t a ~ t point to know various sources o ause of the many trade-offs in aerodynamic drag reduc- tion. This leads to the next point of discussion.

In Figure 16 the relative drag values are compared for a high performance aircraft. Leading the l i s t is the friction drag, with induced drag a close second.

Cross flow or 3-D effects can cuuse drag problems and are unfortunately the most difficult to predict. Induced drag primarily a function of wing aspect ratio can be reduced by wing-tip modifications, as will be covered by Langley Research Center.

Historical Survey o f Drag Figure 17 presents the variation of drag based on wetted area as a function of time. Starting with the Wright Brother's design as the highest drag vehicle--not too surprising if you recall how large a drag penalty wires can create. The lowest drag values correspond to fighter aircraft such as the Douglas A-4 and LTV F-8.

There i s no question that improvements have been made w i t h time, but how well are we doing in realizing the goals of drag previously noted. Shown i n Figure 18 i s a comparison of flight drag data with fiat plate skin fraction curves for turbulent and laminar flow conditions. The data which are for typical general aviation aircraft fall short of even achieving the turbulent flow drag values. The lowest drag value quoted is for the black buzzard (coragyps atratvs) which in some 150 million years of evolution has no doubt managed to achieve reasonably good flow conditions without having to contend with cooling drag and propef ler sf ipstream effects. There are indications that these idealized goals can be approached by aircraft with good surface finishes, such as the point for the Learjet at 30 million Re.

Concluding Remarks In conclusion, three main points should be kept in mind during the next three days (see Figure 19). We need to more accurately clarify the sources of drag for general aviation-type aircraft so that new designs can benefit from more accurate by knowing more about the sources of drag it will be prediction techniques. Next, possible to bring out the greatest potential for drag reduction. Finally, we must use our expertise to identify gaps in knowledge and point out areas which should receive high priority R and D efforts.

A drag ~nforma~ion, 2 10 mph I r ~ ~ r e ~ e ~ t ing a change in equivalent flatplate area from 16.1 to 7.2 square feet.

t

E

B

rc-

.- z

P

e

P

t!

c C

.-

c 7 J a, cu: c

.-

c U

.-

Itr

Yi

; a,

I

Q

f cn

.-

L L

n

I) a 1 )

a

Q J

Q

Q

S rn

.-

U c

b

E ; u- Q, 0 M v)

I

0) Q) h Y In

d .-

P a

s

n z*) Q) L 0,

.-

U

* a

I

0 ti

B s

? D

a U m C

.-

.

& S e, L

.-

P

+

.-

a Q, L .

0)

.-

!A

0 a 0 b b

b,

E !

n

.-- re.

E U .e- o u .

cri

U J

B

n

b

Y c

e

c c

ru

u"

% Y- O Q) Y Lu c Q)

I

E ?

b

3 f

t n c

z

c

P

tr,

Q

z!

0 4- u

e

L L P- c

a

f

Q,

Q .-

LL Kn !!

n

'z

E>

E

\ .-

C

\ i

C v)

\ v)

a , S

\

Kn rr: rc.

\

4 . .

a ,

. \

W

\

hl F

\

3) 0)

.-

u , * m

e

Q

.- 2

I?

Q I cc C c t

.-

E

pt

s

v)

s

Q

W 0

e

*

0 \

Q

m

0 e e

PERS OF SESSION I - STATUS OF

3.

3.1 Overview of Drag Prediction Methods 0 . Ruhmel, Cessna Aircraft Company Prospects and Time Tables for Analytical Estimation of the 3.2 Drag of Complete Aircraft Configurations F . 0, Smetana, North Carolina State University Summary of Drag Cleanup Tests i n the NASA Langley Full- 3.3 Scale Tunnel M. 0. McKinney, NASA Langley Research Center 3.4 Simp1 ified Theoretical Methods for Aerodynamic Design J. Tulinius, NASA Langley Research Center 3 . 5 Drag Reduction/Back to Basics 0. W. Nicks, NASA Langley Research Center 3.1 Overview of Drag Prediction

. Ruhmel

This paper was not submitted for inclusion in these proceedings.

Preceding Page Blank

the Drag of Complete Aircraft Configurations -

ederick 0 e Smetana North Carolina State University The estimation of the aerodynamic drag of a proposed subsonic aircraft config- uration i s s t i l l largely an art practiced with more or less s k i l l by those called upon to perform it. For bodies such as fuselages and nacelles, one usually employs a correlation of wind tunnel and flight test drag data agaimt finess ratio and surface area for generally similar bodies at low angles o f attack as a basis for estimation. Wing and empennage profile drag are usually estimated from the rather extensive collection of airfoil test data which i s now available. The drag due to l i f t can be determined i n what may be called a semi-empirical fashion, that i s to say, an adequate theory i s usually simple enough to apply with perhaps some biasing here and there to make i t agree better with experi- mental results. Interference effects, power effects, cooling losses, and protuberance drag are almost always obtained by extrapolation from previous experience.

The reason for following the aforementioned procedure i s quite simple: It's the only one, one could realistically conceive of undertaking-until recently ct least. Now, however, the situation i s beginning to change, Largely, because of the capacity of the digital computer to carry out literally millions of calculations in- expensively in a short period of time, i t i s now possible to 1 . Determine i n a rigorous fashion from fundamental principles the lift, drag, and pitching moment of airfoils without concave surfaces at moderate angles of attack with good accuracy.

2. Determine reliably the lift, drag (profile as well as induced), and on moderate-to-high aspect ratio unswept pitching moment disttibutions wings or, alternately, the lift, pitching moment, and induced drag only of wings of arbitrary sweep and aspect ratio.

3. Determine with a fair degree of confidence the drag of quasi-streamlined bodies having a plane of symmetry, if the body i s aligned with the stream.

4. Determine in some instances the interference effects of nacelles and fuselages on wing lift, or alternately the inviscid pressure distributions on simple, complete configurations.

Any of these four calcualtions can be done today in less than 15 minutes at a cost of less than $80.00. A more significant expense i s frequently the preparation of an input data set to the computation program. For a fuselage some 1800 coordinates accurately

R Q Preceding Page Blank

representing the half-body and related so as to describe the body surface by qwad- rilaterals of nearly equal area are required.

The boundary layer routines used in these calculations are two-dimensional momentum integral types, although on simple axisymmetric bodies at zero angle of attack as well as airfoil problems, finite difference calculations are possible without exhorbitant additional cost or excessive computer storage requirements. The use o f steady-flow, two-dimensional boundary layer model and i t s associated calculation techniques, however, make i t difficult to locate the flow separation point accurately.

Their use makes i t almost impossible to determine the flow behavior i n the separated wake (where the flow i s almost invariably three-dimensional and unsteady). Flow models and calculation procedures to overcome these deficiencies are known but require computation times and computer storage two-to-three orders of magnitude larger than are presently practical for routine engineering analysis. As a resuft, completely analytical treatments of 1. the lift, drag, and pitching moment of bodies at angle of attack the behavior of airfoils and wings near CL and beyond 2.

max flow separation due to interf,nrence 3.

4. viscous flow over swept and low aspect ratio wings 5. turbulent onset flows containing a helical component and/or energetic streamwise component 6. disturbances produced by protuberances; cannot be anticipated until the necessary computer hardware i s available, estimated by Chapman (Ref. 1) to be about 1985.

There are, however, a number o f developments known io be in progress which should reach fruition by the time the next generation of *'number crunching" computers reaches the market, about 1977. Among these are 1 . Improved singularity distribution techniques which permit the inviscid flow shield over bodies to be calculated with fewer but curved panels and which can give reliable results for bodies having concave surfaces.

Improvements on the Allen-Perkins method of estimating the forces on 2.

inclined bodies of revolution wherein the "inviscid" portion of the flow i s to be calculated from a distribution of singularities.

3. Availability of three-dimensional boundary layer calculation routines for simple but non-axisymmetric bodies.

Availability of optirnirution aigorithims I inked t o viscous flow field 4.

calculation schemes to permit one to specify the aerodynamic behavior of bodies or wings d ired and obtain the g e o ~ e t ~ y which w i l l provide ater, this approach can be expanded to more complex con- fjgurat~ons e I f progress i n computer hardware continues as expected, then by about 1990 i t should be possible to input a contemporary configuration, state some constraints as to performance, stability, and geometry, and the program will produce the geometric offsets for a modified configuration which w i l l satisfy the stated constraints in an optimum fashion. Other programs could then be employed to produce the requisite structural configurations and to drive appropriate numerically-control led manu- facturing equipment. Whether these things come to pass will be dependent upon 1. The cost of developing the programs. Presumably this would be borne largely by the government.

The cost of running the programs. This is largely dependent upon the 2.

availability of hardware of the requisite speed and capacity.

3.

The cost of engineering and technician labor to implement the programs t o do the computation task o r parts of i t manually.

or alternately, The economic incentive to modify an existing aircraft or to build a new 4.

one for improved performance and stability with the same fuel.consumption.

From this vantage point at least, one would estimate that a 10% increase i n development cost could be tolerated if these procedures can yield a significant improvement (10%) in vehicle performance with the same power plant and with no degradation i n handling qualities.

References Chapman, Dean R., "Remarks presented at the NASA Conference on Aero- 1 .

dynamic Analysis," March 1975.

Hicks, R. M., and Vanderplaais, G. N., "Application of Numerical Optimi- 2 .

zation to Airfoil Design," NASA Conference on Aerodynamic Analysis. Also NASA TMX-3213, March 1975, 24 pp.

3.3 Summary of Draa Clean-UD Tests i n 0. McKinney Marion NASA Langley Research Center Introduction Before I start to describe the drag clean-up work in the langley full-scale tunnel, l e t me recognize the pre-eminent work i n the drag field. The late Sighard F. Hoerner i n his book Fluid-Dynamic Drag (Ref. 1 ) has done a wonderful job of pulling together, organizing, and summarizing the vast and fragmented knowledge of aerodynamic drag. His book i s the Bible on the subject. The book is vastly detailed i n i t s presentation and references, and i s about a l l one would need to work the drag problem for general aviation airplanes other than those that are pushing the drag-rise Mach number. Such high-speed aircraft are particularly subject to compressibility and interference problems which will be addressed by Mr. Thomas C. Kelly i n his paper for the wing-drag session of this workshop.

I t should be noted that Hoerner's book i s not a "how-to" book. I t does not set forth a design procedure. Any sensible aerodynamicist knows that airplanes are not designed for low drag alone. They must be designed to do their job (accommodate people, etc .); they must be designed for practical, economical manufacture; and they must be designed with enough sex appeal to sell. So Hoerner's book, i n effect, tells how to get to the ideal shape, md the drag price for departing from that ideal shape.

Thus i t gives the'drag information needed for' trades of performance versus other requirements.

Potential i n Dran Clean-UD Now to get to he specific subject of this paper--the drag clean-up work i n the full-scale tunnel. This work was done between 1935 and 1945 on W. W. I1 fighters and light bombers; so there i; reason to question how applicable i t i s to today's general aviation airplanes. I f i t is applicable at all, i t i s obviously most applicable to the propeller powered airplanes. Since I am not very well acquainted with general aviation airplanes, I started out by making a few calculations that would l e t me see in terms of size, shape, and drag how some of today's general aviation airplanes compare with the small military airplanes that were the subject of the drag clean-up work.

Preceding Page

a t a ~ u ~ a t i o n which resulted from these calculations. Here are compared certain character~sticsof early W. W. II fighters and today's light twin general aviation aircraft. he ~ g u r ~ s given i n each column are averages for five air- hese aircraft did not differ from the mean by more than "-15 percent i n any item. The W. W. I1 fighters selected were the cleaner of the 23 aircraft for which results are summarized in references 2 and 3. There are the P-40, P-41, P-51, P-63, and F 4 F . They are early models, i f not prot from the low gross weight. The light twins are those for which drag could be calculated from information given i n a recent issue of -that is, aircraft for which maximum speed was quoted a t sea level or within the range of altitude for which supercharged engines were flat-rated. The calculated characteristics of the light twins are therefore no better than the information i n "Jane's" or assumptions of 80 percent propulsive efficiency and 80 percent span efficiency factor with regard to drag due to lift.

Now let us trace through the figures and see what we can conclude. The two classes of airplanes have v e r y n e a r l y the same span and length. The light t w i n s hove 20 percent less wing area and 10 percent more wetted area, which factors would tend to cause their drag coefficient to be 30 percent higher than that of the fighters. On the other hand, they have only about one-half as much engine to cool which tends to offset one-third to one-half this difference. So, for equal aerodynamic cteanness we might expect the light twins to have 10 to 20 percent higher drag than the fighters.

The figures show that the value of CD (drag coefficient at zero lift) for the present light twins is indeed slightly higher than that of the fighters as received at the full- scale tunnel before the drag clean-up. Actually since the value of C D ~ for the light twins i s only 10 percent higher than that of the fighters (as received), it would seem that they were slightly cleaner. (&!fore we go further, note that the measured values of CD for the fighters have been corrected for Reynolds number effects from the 80 mph speed a t which the tests were run to the 200 mph speed of the light twins.)

Let us continue by running down the rest of the drag figures for the fighters a s received. The friction drag coefficient i s that calculated on the basis of the wetted area, body and airfoil thickness, and a fully turbulent boundary layer. The cooling and parasite drag values are specifically those o f the P-4? which w i l l be discussed later, The figures for the cleaned up fighters show marked reductions in cooling and parasite drag. These reductions were achieved by reasonable changes which could be made on a practical operating airplane. The friction drag could not be reduced without impractical surface smoothing. But there was Q substantial reduction in total CQ,.

4 4 For the present light twins, the friction.drag was calculated and was larger than hat o f the fighters because o f the greater wetted area and because of the smaller wing area used as a reference. he cooling and parasite drag could not be separated.

last column indicates what the potential for drag clean-up might be. O n the basis of the cleaned-up fighter data, the parasite drag might be similar to that of the fighter, the cooling power of the engines of only one-half the total power might be only one- half as great; and the total C might be reduced from 0.0255 to 0.0195. This i s a DO 13 percent reduction which would result i n a 4 percent increase i n speed, or a 13 per- cent increase i n range or reduction i n fuel consumption at the same speed.

The foregoing figures are admittedly very rough, but they indicate enough potential for drag reduction to warrant pursuing the subject. Another conclusion that might be inferred from the drag figures for the light twin i s that parasite drag can hardly be responsible for the high drag over and above friction drag; so there must be very substantial gains to be made in cooling drag.

Full-Scale Tunnel Tests ~ ~~ ~ The remainder of this paper will examine some of the principal items i n the W. W. I1 airplane drag clean-up work which accounted for considerable amounts of excess drag. And by the way, I have gone over general aviation aircraft i n the NASA hangars, and a t our local airport, and have found all of these items on current general aviation aircraft--not all on any one aircraft, of course. The total of all of them would make the value of C of the average light twin of Figure 1, 30 percent DO higher than that shown.

Drag Clean-up Tests of a Representative Airplane The procedure in the full-scale tunnel tests was to remove all the protuberances from the airplane, to seal a l l openings, and to fair obvious sources of drag such as a blunt sealed radial engine cowling. The drag of this sealed and faired airplane was measured and i f there was any reason to suspect that i t was unduly high, the trouble spots were sought with tufts, surface pressure data, and wake surveys and were then refaired to give a good basic shape. Such a sealed ttnd faired condition for the XP-41 airplane i s indicated in Figure 2.

As the seals and fairings associated with the powerplant installation were re- moved one-by-one, the drag of the following items was identified as show in Figure 3, the drag values being given i n percent of the drag of the airplane in the sealed and faired condition: en engine cowling engine and e x i t 18.6% to permit cooling air flow Unfaired carburetor airscoop 3.6% Cooling airflow through accessory compartment 3.0% Projecting exhause stacts and open hole through which they project 3.6% Intercooler 6.6% Oil cooler 10.2% The total drag of these items associated with the power plant installation increased the drag 45.6 percent above that for the sealed and faired condition.

The drag for the additional features required to bring the airplane t o service condition are shown in Figure 4 by the underlined numbers: 5.4% Removing seals from gaps i n cowling flaps 1.8% Opening case and link ejector chute 1.2% Opening seals around landing gear doors Sanded walkway 4.2% Radio aerials 4.8% Guns and blast tubes 1.8% The total drag of this group of protrusion, roughness, and leakage items equals 19.2 percent of the drag for the sealed and faired condition.

Look a t what h a s happened to the clean airplane we started with! In order f o make it useful we have increased i t s drag nearly 65 percent mostly by adding items that by themselves do not appear particularly large.

All of this drag, however, i s not necessary . Additional tests and careful

analysis showed that the drag of the power plant items could be reduced to 26,6 per- cent and the drag of the roughness and leakage items could be reduced to 2.5 percent, thus saving nearly 36 percent of the drag of the basic condition.

I t i s particularly important to note that in general those items have drags o f only a few percent each. Yet, when taken altogether, they add up to an impressive total.

We started with an airplane i n Figure 1 that was exceptionally clean and i n bringing it to a usable configuration unnecessary drag was added along with the drag associated with the necessary functions. The message here would seem to be that there i s a lot to be gained from attention to details in aerodynamic design.

Design Features Contributing to Excessive Drag The following selected examples illustrate some of the design features i n which lack of attention to detail led to excessive drag.

Cooling drag - The first principles of reducing cooling drag are: do not take

i n too much air, keep the internal flow ges clean, and d e air to the surface i n a streamwise direction. But look, in Figure 5, at what a difference details can make. An exhaust collect or ring, cowling-flap actuating lin a sharp l i p just inside the cowling flap outlet caused an increase in drag of 0.0007 which i s 5 percent as great as the friction drag of the entire Variable cowling outlet flaps are, of course, used to reduce co high speed conditions; but look, in Figure 6, at what leakage through joints i n the flaps can do i f they are not sealed. High pressure air from inside the cowling squirts out normal to the stream causing an increase i n drag of 4 percent of the airplane friction drag for this case. Such cowling leakage was a very common cause of excessive drag in the World War I1 airplane as received at the full-scale tunnel. It could probably be more properly classified as leakage drag than cooling drag, but i n this paper I have chosen in most cases to relate leakage drag to the functional item with which i t is associated.

Engine exhaust stacks - I t would seem that exhaust stacks i f properly recessed

or faired and turned rearward would cause virtually no external drag, but improper treatment of exhaust stacks can result i n large amounts of drag as shown in Figure 7.

The installation shown at the top of the figure appears very similar to the treatment of the exhaust nozzles of turboprop engines i n some of today's general aviation airplanes; and i t caused an increase i n drag corresponding to 16 percent of the friction drag of the entire airplane. The installation shown at the bottom of the figure does not pro- irude into the stream, but caused a drag increase of 8 percent of the frictibn drag because the exhaust gases and the cooling air coming out the hole around the exhaust stacks were ejected almost normal to the airstream, It i s also evident from such installations that designers sometimes fail to take advantage of the considerable thrust the exhaust gases can affort i f directed rearward.

I do not have data for today's general aviation engines, but based on the exhaust gas thrust per horsepower of World War I1 fighter engines, I would expect the thrust coefficient ( ) of the average light twin of Figure 1 to be 0.0026 at full power q i and a speed of 200 knots. This i s enough thrust to offset 15 percent of the friction drag of the airplane.

Landing gears - Even retractable landing gears can have considerable drag

i f not properly treated. Figure 8 shows that the fully faired landing gear shown at the top of the figure had a drag of 7 percent of the friction drag when the seals over the joints were removed. This drag was caused by air leakage through the 1/8-inch cracks around the coverplate. Removal of the rear door to expose half the wheel resulted i n an a d d ~ t ~ o n a ~ small (2 percent C ) increase i n drag. his result, that failure to seal the landing gear doors caused considerable drag, WQS found repeatedly i n the drag clean-up tests.

Control surface gaps - Figure 9 indicates that tail surface gaps can cause an

increase in drag of about 5 percent of the friction drag--and i t would seem thatthe ailerons could cause an additional 2- to 3-percent increase. Such control surface drag can result from several sources. Air can leak through unsealed gaps from the high pressure side of the surface to the low pressure side where i t can exhaust normally to the stream as a jet spoiler. The base drag of the blunt rear of the fin or stabilizer can cause considerable drag, both directlyas base drag and additionally, by pumping air through the airframe if there are lightening holes in the rear spar. Hoerner indicates that such base drag can be reduced markedly, in fact the drag of the entire tail can be reduced 20 percent by reducing the thickness of the airfoil at the blunt base of the fixed nearly 10 percent so that i t i s thinner than the maximum thickness of the control surface about surface.

Irregularities and leakage - Figure 10 shows the results of irregularities and

leakage in one small area of a wing which had a fold joint and a number of access panels. Probably very few general aviation airplanes have features corresponding to the wing-fold joint, but the total number of doors and access panels might be even larger than for this case. I n any event, most drag of this type can be eliminated by better fitting and by elimination of air leakage.

Walkways - Figure 11 shows the drag coefficient of Q sanded walkway to be

0.0010, o r 8 percent as great as the friction drag. This is an extreme case because the walkway protruded abowt l/ll-inch above the wing surface. But, even for more representative cases, the walkway drag was two-thirds this great.

Conclusions I t would seem that two geneml conclusions might be drawn from the foregoing analysis and examples.

1 . There is probably considerable possibility for marked reductions in the

cooling drag of general aviation airplanes with reciprocating engines.

2. Careful attention to detail design and fabrication can result in substantial reductions in drag.

References 1 . erner, Sighard rag, published by the author, 1958.

Library of Congress Caralog Card No. 9 - 1 3009.

Dearborn, C. H., and Silverstein, Abe, Drag Analysis of Single-Engine 2, Miliiury Airplanes Tested in the NACA Full-Scale Wind Tunnel. NACA Wartime Report ACR, October 1948.

3. Lange, Roy H., A Summary of Drag Resulk from Recent Langley Full-Scale- Tunnel Tests of Army and Navy Airplanes. NACA ACR No. L5A30, 1945.

. . . .

m o In

2 2 a

0 0 . 0 . . . .

L s , w f

.-

I u c3 c 2

cr

c

E

Q) c Y

x

c .

/

t

..

Q, t

z

c

.-

c

.-

-u C V 9) f c

.-

d) c ICI a3

-

.u m C e - ;en

-

t

.-

I

-

V rw c

..-

cn . I L t c

.-

E

.cJ Q, u v) Q, v) L w

8"

a '

U cs

du

- z * . , ' : c t $ : , -.c x w h ' L ".

J 0)

E

n , a I I i

i

L

P

-4 -3 c v) OD U c c Q a l a ,

au

. t " .

N S i m d i fied Theoretica Aerodynamic Design Jan R. Tulinius NASA Langley Research Center Introduction this paper is to describe theoretical procedures which can The objective of be utilized by the general aviation industry for aerodynamic design. It i s recog- nized that the general aviation industry has a requirement for a wide range of levels of sophistication in aerodynamic design. However, very simple procedures requiring minimum computer size and operational costs, coupled with minimum input effort and maximum numerical stability, are in general required.

This paper i s organized to first discuss the design process and theoretical methods used to design a wing. Then theoretical methods for estimating the inter- ference velocities due to the fuselage, o r other bodies, and nacelles are discussed.

I t i s assumed that the flow fields due to the different components can be super- imposed, and then the pressure coefficients computed from the Bernoulli equbtion.

Methods to estimate the induced, viscous form, and compressible drags are also discussed.

In addition, a procedure for modifying the surface contours to reduce adverse pressure distributions induced by component interference i s discussed Source/Vortex Lattice In order to theoretically design a finite aspect ratio wing, it i s necessary to have a thick-lifting-surface theory. If the boundary conditions are linearized, the flow fields induced by the wing can be divided into two parts: (1) the flow field due to lift; and (2) the flow field due to thickness. Also, due to the Iineari- lift can be linearly related to the local zution, the perturbation velocity due to angle of attack of the mean camber surface and the perturbation velocity due to thickness can be linearly related to the freestream component of the gradient o f the thickness distribution.

The most simplified thick-lifting-surface theory i s the source/vortex lattice shown in Figure 1. The vortex lattice i s a system of lifting lines where one lifting line is placed on each quadrilateral panel of the wing. The source lattice i s analogous to the lifting line, except that the velocities induced by a source line are rotat the vortex lines cannot end

, and, therefore, there ng off to infinity. The

laced a t the quarter chord of each panel and the source nd three-quarter chord points of each panel The wing surface velocities are computed at the three-quarter-chord point. Derivations of the influence equations for the source/vortex lattice along with second ord f o r blunt leading edges, interference between thickness and lift, a rections compressibility are given in References I and 2.

The source/vortex lattice can be used to solve for the surface pressures for a given wing twist, camber, angle of attack, and thickness distribution, or it can be used to solve for the wing twist, camber, angle of attack, and thickness distri- bution for a desired set of surface pressures.

Figure 2 describes how the wing can be designed using the source/vortex lattice. If an airfoil has been designed with the desired section properties, the wing surface shape, which w i l l produce the same upper and lower surface presswres as the two-dimensional airfoil, can be solved for using the approach shown on the left side of figure 2.

If no two-dimensional design i s available, with the desired section charac- teristics, the second approach shown on the right side of Figure 2 can be used.

This approach i s analogous to the two-dimensional "ideal angle of attack" design technique discussed i n Reference 3. Essentially what this design procedure does i s define a wing which has minimum induced drag at the design lift coefficient and a minimum adverse leading-edge pressure gradient, for low viscous form drag, at what i s defined as QOpT; in Figure 2.

As shown in Figure 2, this design procedure gives the wing camber and twist such that the induced-drag polar is tangent to the locus of minimum induced drags at C L ~ and the zero percent suction polar i s tangent to the induced-drag

polar at C L ~ ~ , . . The data should fall somewhere between these two polars. The

.

zero percent suction drag i s defined as that for which there i s no thrust at the leading edge of the section.

General Slender Body Theory General slender body theory can be used to compute the flow fields due to arbitrary shaped fuselages or other bodies. The theory requires access to only a body of revolution and a two-dimensional airfoil theory. The arbitrary body flow fields are computed, as depicted in Figure 3, by superimposing the flow fields from the e~uivalent body of revolution' plus a correction for the noncircu ar cross-section, which i s computed with a ~ o - d ~ m e n s ~ o n a l airfoil program, he correction flow field due to the noncircular cross-section i s obtained by subtracting the iwo-dimensional solution for the circular cross-section from that for the noncircular cross-section. These solutions are obtained using the ful I three- dimensional boundary conditions in the two-dimensional airfoil program. The difference between the usual two-dimensional boundary conditions, which would represent the effects due to angle of attack and sideslip, and the three-dimensional boundary conditions is the effect due to body growth in the longitudinal direction.

There are several possible variations to the solution of the general slender body problem. Langley Research Center i s in the process of developing a version which will be coupled with a thick-lifting-surface theory. A derivation of the theory is given in Reference 4.

The application of this theory and i t s component parts are l i s t e d in Figure 4.

Induced-Drag Analysis and Design The induced drag o f an arbitrary lifting system can be computed by an equivalent lifting line in the Trefftz pland. Figure 5 depicts the equivalent lifting line as seen from an end view. The dots represent the trailing legs of the horse- shoe vortices pointing out of the paper. The bound vortex line segments between

the dots have circulation strengths equal to r. The line can be bent to represent

any arbitrary wake shape. Also, any number of wakes can be represented to account for multiple lifting surfaces such as horizontal tails, canards, fins, and vertical tails US well os the wing. End plates or winglets can also be represented by this method.

As listed in Figure 6, the theory can be used to solve for not only the drag of a given configuration, after the span loads have been computed by the source/ vortex lattice theory, but due to the quadratic nature of the drag expression, the optimum span loads can also be computed for both trimmed and untrimmed conditions using the method of Lagrange mu1tip1 iers.

A computer program has been developed at Langley Research Center by Blair Gloss and the author to compute the induced drag for given span loads or solve for optimum span loads utilizing the equation in Figure 5 . This equation i s derived in Reference 1.

Viscous Form Drag The viscous form drag I s due to the boundary layer giving the airfoil section an effective cha ge in shape due to Phe djsp~acement thickness. As isted in Figure 7, the viscous form drag can be estimated from two-dimensional experimental drag polars at each span station along the wing, utilizing the wing section lift coefficient, and then integrating. T h i s approach is presently being developed by Professor Ralph Krenkel of the Polytechnic Institute of New York under a NASA grant from Langley Researcb Center.

The viscous form drag can also be computed using an infinite yawed wing boundary-layer program at a series of span stations and then integrating. This approach i s being presently worked on at Langley utilizing a boundary-layer program developed by Frank Dvorak and Frank Woodward (Ref. 5). I n this approach, equiva- lent airfoils are developed a t eoch span station which produce the same pressure distributions in two-dimensional flow a s i s developed by the actual section on the finite wing in three-dimensional flow, These airfoils are then run through the infinite yawed wing boundary-layer program to determine the section viscous form drag.

A third procedure, which i s most appropriate during preliminary design, i s to utilize a percent suction versus C curve from a configuration with similar sec- L tion properties. This curve, shown in Figure 8, i s obtained by the following equation: % suction = x 100 2/ 2/ The two boundary curves defined by CL CL, and CL TAR are the upper and lower bounds, respectively. The location of the data relative to these two curves i s primarily a function of viscous effects due to the section shape. As can be seen i n Figure 8 , the GAr/V)-l section has good characteristics up to a CL = 1.2. The theory shown on this figure i s from a vortex-lattice program developed by the author.

Compressible Drag The cmpressible drag for conventional airfoils can be estimated using the crest theory. The crest theory states that drag divergence w i l l occur shortly after the crest of the airfoil becomes sonic. The crest i s defined as that point on the airfoil where the free stream i s tangent to the airfoil surface. The critical pressure coefficient is define on Figure 9 . it should be noted that it i s a function of both sweep and the free-stream Once the Mach number for drag divergence MDD i s known, then an incre- mental drag due to compressibility can be obtained from an empirical curve of A Q C versus M/MDD. T h i s increment i s then added to the incompressible drag obtained from the sum of the skin friction drag, induced drag, and viscous form drag.

Contour Modifications due to Interference Effects With the procedures discussed in the previous sections, the surface pressures and associated pressure drag and lift can be estimated for a complete aircraft.

Since the wing alone was designed to have an optimum pressure distribution, the addition of the fuselage, tip tanks, and nacelles will deteriorate the wing-alone design. These interference effects can be minimized by modifying the component contours to relieve the unfavorable interference pressures.

The incremental pressures due to the interference can be converted to incremental velocities through the use of fhe Bernoulli equation. As much of this increment as possible should be relieved by judicious placement of the components.

Then the remainder should be minimized by locally contouring adjacent components.

in wing shape to account for flow induced by another component The change can be solved for by means of the relationship between velocity and the slope of the surface given in Figure 1. If the induced velocity i s primarily in the chord direc- tion, the thickness can be modified by the process outlined in Figure 10.

The contour modification might have to be divided between the wing thick- ness and the adjacent component surface. If the adjacent component is a body, approximately twice as big a slope change must be made for the same change in velocity as i s needed on the wing. T h i s i s due to the fact that the perturbation velocity produced by a two-dimensional contour i s about twice as large as that produced by a body. This i s only an approximate rule of thumb. However, for the case of a sphere and a cylinder, the difference i s exactly twice.

If the induced flow is perpendicular to the chord, then a twist and camber modification will be needed. In this case, i t i s just necessary to change the mean camber surface angle of attack by an amount equal to the induced angle of attack.

There exist sufficient simp ified theories applicable to the design of general aviation aircraft. However, most of these theories have not been programmed for general aviation design purposes, and in many cases, the computer programs and their application have not been published in the open literature.

Langley Research Center i s supporting the development of programs, utilizing these theories, which will be sized for general aviation purposes. Some of these programs are being developed by center researchers and others under university grants. A significant portion of this capability will be available a year from now.

References 1. Tulinius, J., Clever, W., Niemann, A., Dunn, K . , and Gaither, B . , "Theoretical Prediction of Airplane Stability Derivatives at Subcritical Speeds, " NASA CR- 132681, 1 975 e Tulinius, J., "Theoretical Prediction of Thick Wing and Pylon-Fuselage- 2 .

Fanpod-Nacelle Aerodynamic Characteristics at Subcritical Speeds, " NASA CR- 137578, 1 974.

3 . Abbott, I., and von Doenhoff, A., Theory of Wing Sections, Dover Pub1 ications, Inc . , 1959.

Ashley, H . , and Landahl, M., Aerodynamics of Wings and Bodies, Addison- 4.

Wesley Publishing Co., Inc., 1965.

Dvorak, F . , and Woodward, F . , "A Viscous/Potential Flow Interaction 5 .

Analysis Method for Mufti-Element Infinite Swept Wings, " NASA CR-2476, 1 974.

> - ' I'B >

N N ul ul U 8 a 0 N N rl CL 0 u 8 8 e "". u 8 - c I I

*

n a x L .

a, L c

+ +

x I- L LLJ . ..

a,

F ? g

a,

-

ln c) a, L CD

.-

u , fl N I Q c .

a U x L .

c O m I-- 4) x u I- x T 3 m L aa W -0 c

%

aa

-

'r vt LL

-

L A

e

.c.

Qt m c c

.- >-

-

I - c3 F- U L1L 0 0 cn m .cJ t

CL .-

U i-7 r : 2!

3 a3

a3 ol 0 -L J= t v) M (0 v) d + w

c

v) d L Q

J >

Q> w c CI I - L IC,

Q -

v) o i 0 d w I

5 ! i

CL m w t

3 3

c3 m

<

).

a

e

e

L3

a

w U C J M 0 w a v)

z

- a

I

-

I -

-

D P

2 -

w 3.

*

N w C Y k - 3 - u f pz

z

i

L L c C L v) a 2

*

L u , c (-..

d a

* z

c3 0 bl-

z -

c 0- c U N

.-

-

c3

z

w

c- w 3

J = € i Q

*

e e

E

€ -

z

ocr

u

J

<

-

c-

v)

e

d z

e

I - v) n

A

c3 e i- cli 0 ai A W a U a VI

.-

k v) _.

>t-

z

e

I-

a

c3

E

M U VI

a

0 .

2 2

*

LL

2%

n o

-

I

-

P-4 I

x

E

e

I

*

OQ

e t - - V ).

n M at > Q)

.-

I -

-

Q) p?: P cv C

T

u

.-

.I- W 0 V

> .-

!?z i- U

a

c3 L w

z

$

C ).

1;

s

e a

e h ,

e a c

e

7 6

uction - Back to

Oran W. Nicks NASA Langley Research Center

Introduction - PersPective

From the beginning of manned flight, the iteration of lift, weight, drag, and thrust have been the balancing factors involved in so-called aeronautical engineering.

L i f t greater than weight i s needed to get up, thrust greater than drag i s needed to go These fundamentals are s t i l l as true as ever. The list of variables involved anywhere.

i n successful aeronautical engineering has grown significantly to include speed, cost, comfort, aesthetics, pollution, noise, etc., with perhaps the most significant current interest in fuel economy. There will surely be other tradeoff's to be faced, but never will we be able to ignore lift, weight, drag, and thrust.

I n this conference, we will deliberately focus attention on drag reduction. Drag i s the basic parameter affecting the ability of aircraft to go somewhere efficiently. A hot gas balloon can get up and stay up reasonably well, with essentially no consideration I t goes when the wind blows, a t no more than the speed of wind. But as soon for drag.

as you decide to make i t go faster than the wind, or i n another direction, its drag becomes very important.

I n the early days, airplanes were a lot like the free balloon--getting up and staying up was difficult enough without worrying much about going somewhere efficiently.

The structures guys were hard pressed to make lightweight structures, and the aero- dynamicists struggled to develop the lift necessary to keep them up. As t k aero- dynamicists really got to working on the drag problem, the propulsion guys came along and helped solve the problem by providing better engines and propellers-that may be one reason we have some unanswered questions about the science of low speed flight today. I often wonder what a few more years o f studying the birds might have produced, had the propeller not allowed an effective alternate to the aerodynamic- propulsion techniques s t i l l employed by the birds.

At any rate, these are the kind of questions I think we should consider during this Drag Reduction Conference, as we look back to basics.

Wi ng Lift- Drag Re latio n shi ps In addition to providing almost a l l the lift, the wing produces the bi,ggest percentage of the drag, a ut 50 to 60 percent during cruise for usual configurations.

ight, getting the required lift with the ce the wing i s ~ n d a m e n ~ l to lifting the least drag has been the challenge for wing design through the years. I f an airplane had some way of getting to cruise speed and altitude, the wing required for cruise might be roughly half the area of the wing required for acceptable takeoff, climb, and landing.

Of course, under such ideal conditions, there would be little need for meiry. In this case, the wing designed only for cruise flight for a four cruising at 200 miles per hour would contribute only about 30 percent of the drag.

To give the same airplane a good takeoff, climb, and landing capability with a plain wing of the same design, the wing would contribute about 70 percent of the total drag at 200 miles per hour. Of course, i t i s that situation which has led to the development of variable geometry high l i f t devices such as flaps and slats. With today's technologies, such devices reduce the wing drag penalty during cruise to about 50 percent of the total; however, there are several basic shortcomings of these devices which we might well consider.

First of all, the most common trailing edge flaps cause increased pitching moments which require increased down loads on the tail for trim. In a typical landing of the wing i s negated by the down load on configuration, about ten percent of the l i f t In addition, high performance flaps which increase the the tail required for trim.

area usually decrease the span efficiency with an associated penalty due to wing if we had variable camber devices or higher induced drag. I t would be helpful variable span techniques to increase lift coefficient while keeping 'the center of pressure forward, and to minimize induced drag a t high lift conditions. Birds use forward sweep, variable camber, variable aspect ratio and lifting tails every day.

Such variable geometry features are tough to design and build; however, some of the newer technologies may make them more attractive possibilities than they have been in the past. The thicker wing section, for example, gives structural depth; new composite materials may simplify controlled bending o f aerodynamic surfaces. While I am not proposing any particular solution to the problem, I do suggest that a thorough review of the basics which cause drag, and some imaginative consideration of techniques for reducing drag, may be productive.

Profile Drag The resistance of an object moving through air i s pretty clearly a function of the cross-sectional area, the wetted area, the shape of the object, and the friction caused by the scrubbing of the air over the object. Here again, the wing, although s t r ~ ~ i ~ n e d i n appearance, contri Utes 20-40 percent of the here i s only s o much that can be done about the cross-sectional area associ- ated with the volume required for passengers o r payload, but there are many other smaller factors which add up in the profile drag account.

Sometimes the quantity and types of protrusions on modern general aviation aircraft make me think it would be helpful if aerodynamicists some simple experiments on stream1ining My teen-age son recently conducted a science experiment in a small wind tunnel to show the effects of streamlining by comparing a circular flat plate, CI sphere, and a streamlined shape with a fineness ratio o f 3 1/2, all having the same diameter. The difference in drag for the plate and the streamlined body i s a factor of about 30, i n case you don't remember. My son's teacher could not beiieve the measurements when the much larger body produced the dramatic reduction, and i think many aerodynamicists would be impressed as well. (I guess I was, even though I knew Hoerner's data were - to be trusted. ) Many airplanes flying today pay a large price in parasite drag for fixed landing gear, steps, antennae, windshields, and the usual joints, rivet heads, doors, and other discontinuities attributed to production. Hoerner has data on German tests of an actual ME-109 wing and on a section of a P-51 wing-both of course being real con- struction though quite different i n detail. The data are not presented i n a manner such that they can be compared over a range of conditions--they are single points-- but they show the drag of the ME-109 wing to be 70 percent higher than that of the P-51 wing. According to Hoetner, the high drag of the 109 wing i s due largely to manufacturing features: surface waviness, holes, cover plates, control gaps, il 1- fitting slat, rivet heads, and bolt heads, whereas the P-51 wing was flush riveted, filled, sanded and painted. The desirability of achieving laminar flow was the motivation for the attention to smoothness on the P-51 wing, although it i s doubtful I t i s likely that the elimination of protuberances that very much laminar flow existed.

helped make most o f the difference. Better fabrication techniques, or possibly surface coverings that might cover up production artifacts, may be worth more attention than they have been given i n the production of many current airplanes. The possibility of applying a space age material coating over a standard production surface i s being studied a t Langley.

' Propu I s io n Drag Another form of drag many of us have gotten used to i s that associated with internal flows aroun engines and accessories. To be sure, the matter is an inter- disciplinary p r o ~ ~ e m involving interfaces with the engine, propeller, and airframe.

ntil the jet engine came along and caused more aerodynamicists to concern them- selves with internal flows, the matter of engine-nacelle drag was largely an empirical NACA's experimental work on cowlings i n the 2 0 ' s and or experimental matter. The 3 0 ' s provided data for use with radial and in-line engines used extensively during the 1940's. The advent of the horizonally opposed engine brought wit tunities for better streamlining and while there are many good examples flying, I am not aware of a systematic set of data on the subject relevant to aircraft and engines of today. Considering the fact that recent workshops have indicated that from 5 to 25 percent of the tatal aircraft drag may be caused by cooling air flows, and knowing ' that the velocity inside cowlings may be well above 100 miles an hour, i t i s clear that drag reduction possibilities exist for future designs i f attention i s paid to internal flows.

From the standpoint of basics, a subsonic ramjet can be made to produce internal thrust with efficient heat addition to air flow. Assuming that the external drag o f a cowl i s a part of the airplane drag, the fact that the engine i s adding heat energy i s .

significant. To take advantage of this, the internal flows and the cooling flow ex- haust must be treated carefully to reduce losses and to recover the air momentum along the thrust axis. While first priority for cooling air i s obviously to cool the engine, there i s nothing whieh says the design should not capitalize on the heat addition. Efficient baffling designs which preclude dumping of air, high speed flows past siructure, supports and other drag producing items, may help make the most of the cooling air situation.

A simple calculation based on data from Hoerner indicates that for a flight speed of 200 MPH, a cooling air flow receiving a 300 degree temperature rise through a 2-sq.

ft. cowl would produce an internal thrust of about 25 pounds. By contrast, a cold engine would produce about 50 pounds of drag. A classic example of turning such potential losses into a gain was the design of the P-51 Mustang glycol radiator, which reportedly produced a net thrust for the complete instal lation.

Propellers have evolved in the face of many compromises, but their efficiencies continue to suffer because of basic tradeoffs. ,The propellers developed by the Wright Brothers provided an ideal efficiency of 80 percent and actually delivered 66 percent of the power available to the airstream. This was achieved by careful attention to theory and the fact that they were large and rotated a t relatively low speeds. As engines have become smaller, their speeds have become higher and the unfavorable gearing have led toward smaf I diameter, high speed propellers. Propeller efficiency i s not only compromised somewhat, but the higher velocity scrubbing and outward flows around nacelles may con ibute ad~itional losses of a few percent.

ost of the general aviation jet aircraft benefit fronr the aft engine locations which tend to accelerate flows near the base of the airplane where wakes and boundury layers are pronounced. Rear mounted propellers have the same potential for flow improvements around the wing and fuselage, but of course, they are not as readily adapted to airplanes as the jet engine and have not found as much use.

Some Anomalies from the Past and Present Since most of the ideas that occur to me have already been exercised i n the past, I make i t a practice to look back frequently to anomalies which may provide lessons for today. Many good ideas have failed to materialize into practice because of shortcomings i n the technologies other than the disciplines being explored. What I am referring to i s the fact that an aerodynamicist with a good idea may have been thwarted because of a structures or materials problem, for ex.ample, and advances i n these other fields may have opened the door later without his knowing it. With this philosophy i n mind, let me challenge your thinking a b i t with some questions which arose out of looking back.

I n the 1930's there was a lot of effort applied to the matter of drag reduction.

This era produced airplanes like the Cessna Airmaster, the Lockheed Vega, and the Northrop Gamma. Larry Loftin, who has gathered data from many sources and done o f the minimum drag coefficients for these many calculations, provided estimates I averaged to be about 0.0270. Similar calculations for representative examples which fixed gear monoplanes of today give an average of about 0.0370. I realize I am comparing the very best of the 3 0 ' s with the average of today, but the question is, "What was i t about those airplanes that made them appear to be better from a drag standpoint that might teach us something." You will have to decide, but l e t me mention a few things to stimulate your thinking.

First, the airplanes considered from the 3 0 ' s were all tail draggers and the current examples considered all have nose wheels. Obviously, some penalty i s being paid for nose gear; k e r n e r gives numbers ranging from 6 to 12 percent, not including effects on propeller efficiency, but this alone does not account for the difference.

Another characteristic of these airplanes of the 3 0 ' s was a carefully cowled radial engine, whereas the examples of today a l l have horizontally opposed engines. The high performance airplanes of the 3 0 ' s were also extremely smooth, usually employing many coats of dope over fabric or plywood and having few extrernal protuberances

(before transponders , s and the l i k e were required). Frontal areas were reduced

to a minimum, and careful attention paid to cross-sectional areas, fuselage shapes, wing taper, win tips, and fairings.

The story for retractables i s somewhat different. Some of the current retractable gear configurations compare favorably with the best of World War II and i t appears that when paying the price of retractable gear, aerodynamicists are also concerned about other forms of drag. However, I do not suggest that you immediately concl that doing as well as World War I1 aircraft i s acceptable for today--that is not a good assumption.

Some more recent anomalies which are of interest because of their concepts should be mentioned. Gus Raspet and his colleagues at Mississippi State University worked hord to reduce drag. Gus recognized the importance of skin friction and did many experiments on methods of reducing it. Some work a t MSU in 1967, not long after his accident, involved several research aircraft with highly advanced technologies.

The XV-11A developed for the U.S. Army had a variable camber fiberglas wing with boundary layer suction and a pusher shrouded propeller. While only 40 hours of flight tests were conducted, significant indications of improvements were achieved. For s t a l l speed was decreased from about 75 knots to about 52 knob with no example, change i n wing area.

All of us are familiar with the work of J i m Bede--I think i t is fair to say that his primary aerodynamic emphasis i s on drag reduction. Aknast every basic principle we have discussed has been considered by Jim i n his decisions.

Another drag reduction effort o f the last few years that impressed me was the work of Wil Schuemann. Starting with a Libelle high performance fiberglas sailplane, with an advertised L/D of about 39 a t 59 MPH, W i l substantially improved i t s high speed capabilities without compromise to i t s low speed performance. His efforts in- volved a combination of improving the flow over the airfoil by modest leading edge modifications and by employing the basics of Dr. Hoerner to fillets, air leaks, control surfaces, and laminar flow surface considerations. His tests show that tk cruise l i f t drag ratio a t 100 knots was improved by 30 percent to a value of about 20; the fact that this was possible starting with an extremely clean configuration illustrates the possibilities for drag reduction by applying basic principles.

While talking about sailplanes and anomalies, i t seems appropriate to note that there are many high performance sailplanes regularly employing the advantages of significant runs of laminar flow. While practical means have not yet been demonstrated For achieving these benefits on higher speed aircraft, there i s a lot of pay dirt for drag reduction between fully turbulent and fully laminar flow. Considering the fact that friction drag accounts for 60-70 percent of the average airplane drag, work on com- pliant surfaces, boundary layer control, and other means of reducing skin friction are "musts" for research.

Summary In a sweeping manner, and with some academic liberties, I have touched on many i t e m that are on the agenda for this conference. Most of you are familiar with the basics, but you may not have had the opportunity to ponder them in one sitting for a long time. I am mindful that this country makes the best general aviation air- in the world--last year, about 15,000 of them. This is a fact for which we can craft be proud, and yet, we are gathered here to discuss the important matter of making them better.

I hope that while your meal was settling, our brief revisit to fundamentals has helped stimulate an open-minded consideration of old ideas with a new twist, Summed up, my comments were aimed a t making two points: (1) We can enhance our funda- mental knowledge i f we carefully review and reconsider the basics unravelled by those who preceded us; and (2) With a good understanding of the basics, we must be as bold and innovative as the pioneers of the past in searching for means of applying new technologies.

4, €RS OF SESSION I - FUSELAGE DRAG

4.1 Overview of Fuselage Drag J. Roskam, University of Kansas 4.2 Propeller Blockage Research Needs R. Tuml inson, Beech Aircraft Corporation Preservation of Wing Leading Edge Suction a t the Plane of 4.3 Symmetry as a Factor in Wing-Fuselage Design E . E . Larrabee, Massachusetts Institute of Technology Asymptotic Analytical Methods in Fluid Mechanics Related 4.4 to Drag Prediction G. R. Inger, Virginia Polytechnic Institute The Economic Impact of Drag in General Aviation 4.5 R. D. Neal, Gates Learjet Corporation

Preceding Page Blank

. I Some Comments on Fuse Jan Roskam University of Kansas Introduction This paper focuses on the following areas relating to fuselage drag:

Fuselage fineness - ratio and why and how t h i s can be selected during

1.

preliminary design; 2. Windshield drag; 3. Skin roughness; and 4. Research needs in the area of fuselage drag Fuselage Fineness Ratio and How It Can Be Selected Table 1 presents some data on fuselage fineness ratios for several current general aviation airplanes. It is interesting to note, that with one exception, all have values of around R$d = 5 to 6. I n Reference 1, the fuselage (or body) drag i s estimated from: A This equation assumes zero base drag. Figure 1 shows how the [ e ]-term i n equation (1) i s related to R$d. Note that the [ ]-term no longer decreases significantly significantly after R $d = 6.0 is exceeded. This would indeed suggest that values of 5 to 6 for R e/d are about optimum. However, there are three other factors to contend with: increasing it$d will decrease Cfg ; 2. increasing R d d will increase Swet ; and body 3.

increasing &$d will decrease tail wetted area requirements, for constant stability levels.

It appears that a more detailed examination'of fuselage fineness ratio i s there- fore in order. The next section presents a method for minimizing the sum of fuselage and empennage friction drag, under a constant directional and longitudinal stability constraint

Preceding Page Blank

+ Z G

-

t n N + 2 4 6 4 3 IO 12

BODY F I N E N E S S RATIO - ‘S/d

Figure 1 . Body Zero-Lift Drag Factor as a Fwnction of Body Fineness Ratio Table 1 . Examples of Fuselage Fineness Ratios and Wetted Areas for General Aviation Aircraft &B Swing Swet %et - body d Swing

I Cessna 210

5.02 175 319 1 . 8 2 Cessna 207 5 . 6 9 174 425 2 . 4 4 5.22 Beech Sierra 146 332 2 . 2 7 5 . 1 5 Cessna 176 292 1.68 Beech Bonanza (‘58 4.98 iai 323 1.78 Beech Baron 5 . 6 9 199.2 362 1 . 8 2 5.97 Piper Navajo 229 502 2 . 1 9 Cessna 310 5.40 179 306 1.71 Piper Seneca 5 . 6 8 206.5 . 356 1 . 7 2 Beech Duke 5.59 212.9 586 2.28 Cessna 414 5 . 5 2 195.7 488 2.49 Beech King Air 6 . 0 6 294 652 2.22 Gates Learjet 24 8 . 8 t 232 502 2 . 1 6 empennage friction drag can be estimated under constant static stability constraints e It is assumed that the fuselage from nose to passenger compartment i s defined roughly as in Figure 2.

Figure 2. Definition of Fuselage in Two Parts It is also assumed that the tail cone can be represented by a skewed cone as in Figure 3.

Figure 3 . Modeling Aft Fuselage as a Skewed Cone The equivalent fuselage diameter is defined such that: The wetted area of the fuselage can now be written as:

L c +

- c

-4" s a correction factor a ~ c o u n t ~ n g for the fact that the rear fuse found by compar~$on to existing aircraft e rag coefficient (zero-l ift) can All symbols are defined in Reference 1. Fuselage base drag i s neglected.

For given , can thus be computed as a function of Rc.

cDofus Empennage Drag - The horizontal tail wetted area may be approximated by: where the geometry i s defined in Figure 4.

d

Figure 4. Horizontal Tail in Relation to Fuselage Cone The horizontal tail drag coefficient can be written as: where all symbols are defined in Reference 1.

The - vertical tail wetted area may be approximated by: where the geometry i s defined i n Figure 5^.

Figure 5 . Vertical Tail in Relation to Fuselage Cone The vertical tail drag coefficient can be written as:

where a l l symbols are defined in Reference 1 . Horizontal and vertical tail sizes

are here assumed to be determined by minimum stability requirements, i .e.,:

Directional Stability - Neglecting the wing contribution, the directional

stability of an airplane can be written a s : where the symbols are defined in Reference 2. The geometry i s defined in Figure 6 .

Figure 6. Fuselage Geometry for Estimating Directional Stability Note than K N and K R ~ are functions of R , . Body s i d e area, S b can be expressed a: where F i s os in equation (3).

Note that:

and 3t 4 (LG rALEy,CRw) QS ilhstrated in Figure 7. (12)

Figure 7. Definition of R for Swept Vertical Tail From the sketch the following equations may be deduced: 2 s v

Cu,=

z , =

Now, substitute equation (13) into (9) while using equations (14), (13, and (16): For preselected values of %,

s , , , b 3 gw

A,, x', and A,,, 3

it is now possible to solve for Sv for any given value of R,.

as a function of 1 Having done that, it i s possible t o compute C C ' O v .'I.

where all symbols are defined in Reference 3 and where: _.

X P L H - - X a c w 3 . 4

as shown in Figure 8.

Figure 8. Definition of Horizontal Tail in Relation to Fuselage

-

It ..L is assumed, that % c W s C L ~ ~ ~ > - c , 3Yd4 > S-+.!\

and x,,, are known and fixed quantities.

The following expressions can be shown to hold:

-

-eH= (1N-X,)s(e,-cRH) +yc U -t;R*ALE,t

G

+ - k l

(21 1

-

c, =: $eR"

(22 1

uation (20) and using equations (22) through (25) i t is found that: Now, setting dC,,/dCL = some constant value and preselecting: AH, A H and A LEH , i t i s possible to solve for SH (using equation (1 9) for any given value of R , .

Having done this, it i s possible to compute CD as a function of Rc).

OHT

Parametric Study - The methods of of the previous sections allow the

computation of CD , C and CD for given values of A

LE(H,V) Ofus ')Oh ,t 0v.t.

and for given values of llc.

These contributions can be plotted against gC/d as shown in Figure 9.

Figure 9. Plotted Results of Parametric Study If need be this process can be repeated for a variety of empennage sweep for minimum fuselage plus empennage drag can angles. The rear fuselage length P , C 9 .

be readily found from Figure It would be of interest to include the effect of weight in this parametric study * shows some resu culations using a Beech King Air as example. It is seen that th ge plus empennage drag i s not tim mum from this point of view e It would be of interest to extend this a t o other airplanes.

. ._ BO6 4 0 4 _. .

.OOj - Figure 9A. Effect of Tailcone Length on Fusela e Plus Empennage Zero Lift Drag Under Constant Stability tonstmints Win& Drag eference 4 presents a series of systematic data for windshield drag of small anes, It sum~arizes by stating that windshield drag can range from 20 to 1 percent of airplane drag depending on how well they are faired.

This i s a wide drag range!

Figures 10 and 11 illustrate the types of windshields investigated i n reference 2.

Figure 12 illustrates a range of windshields found on current general aviation airplanes. It i s seen that windshields of 1975 are quite different from those that prevailed in 1942. It would seem that some systematic research into this area would pay off for certain airplanes.

Surface Finish The subject of skin waiviness and surface finish has not been brought up, because of the strong interplay with production and tooling costs. However, as shown in Figure 13 there i s probably considerable room for improvement. This could be attained by a more wide spread use of metal bonding in aircraft fabrication. This way, it i s feasible to maintain large areas of laminar flow over the forward part the fuselage and capitalize on the resulting lower friction drag.

of Research Needs 30 to 50 percent of total airplane drag.

The fuselage typically accounts for It seems that improvements of at least 10-20 percent could he made by taking a good research look at: 1 . fuselage fineness ratio; 2. windshield drag; and 3. low cost application of metal bonding to reduce skin friction drag.

It would seem that research in the area of windshield drag should be in the form of a series of systematic wind tunnel tests.

Optimization of fuselage fineness ratio couJd be achieved through the development of an appropriate computer program which would also account for the effect of weight.

M A L A REPORT NO. 730 ' I ' 4 .

., I

Sin le-cwved Doubfe - cwved

&s widshield, glass windshield,

, ShWP , d R o d ec4Fe. edge (3

[a) Id) -.- -----I- V 0 - Singfe-curved Fooired nose glass windshield, round edge Figure 10. Drag of Fuselage with Transport-Type Windshields M, 0.35; V, 265 mph NACA T U . .

( I -!-a),

n o I

9-- ..

I

. mfi

(l-l.-3)d .

V-- I - - - - - -

* - a: ,AI

+------ .I60 . t40 . IZO * Q % * C.lo0 . u f .

e .on0

9"

.OS0 .mo .oco ?< ,Fuseloge angle of otiock. d , , deg , Figure l l a . Effect of Retaining Strips Figure l l b . Effect of Retaining Strips, Combination 1-1-3, M, 0.34; V, 260 Combinations 1-1-3 and 3-1-3, M, 0 . 3 4 ; V, 260 mph mPh

fa

Beech King A i r AI00 G r u n m a n American AA-5 Traveler Plper Cherokee Warrior Gates Leai-jet 24D Cessna Cardinal RG Beech Duke B60 Cessna Skywagon 207 ' Figure 12, Typical General Aviation Windshields for 1975 N A C A T R q \ O Figure 13. Effect of Surface Improvements on Drag Characteristics of Airfoil Sections eferences

2 . Roskam, J .; Methods for Establishing Stability and Control Derivatives of

Conventional Subsonic Airplanes; Pub1 ished by Roskam Aviation and Engineering Corporation; 519 Boulder, Lawrence, Kansas 66044.

Roskam, J.; Flight Dynamics of Rigid and Elastic Airplanes; Published 3.

by Roskam Aviation and Engineering Corporation; 519 Boulder, Lawrence, Kansas 66044.

Robinson, R.G. and Delano, J.B.; An Investigation of the Drag of 4.

Windshields in the 8-foot High Speed Wind Tunnel; NACA TR 730, 1942.

1 02 lockage Research Needs R. R. TumIimon Beech Aircraft Corp.

If the general aviation industry i s to produce the most efficient airplanes, it i s important that the best technical tools which can be economically used be employed. That is the business of this workshop. One of the technical toois that is needed and which is currently not available i s a means to accur propeller blockage.

Propeller blockage refers to the effect of mutual propeller-nacelle or

fuselage interference on the propulsive efficiency . The interference of the body on

the propeller arises from the retardation of the airflow through the propeller disk and the resulting change in advance ratios. The interference of the propeller on the body stems from additional drag on the body because of the slipstream effect on local pressure and boundary layer. This effect has been understood for many years, and there are many reports in the literature. In the interest of brevity, the present body of information wilt not be explored in this presentation except for the biblio- graphy included and to note the important paramdters of advance ratio, body shape, and the propel ler-diameter-body-diameter rat io.

However, while the sources of the propeller body interference have been understood for some time, the experimental data available to allow accurate estima- tion i s long out of date. Current configurations with horizontally opposed engines outdate the data available which was determined with radial engine and in-line engine configurations as well as RAF -6 and Clark -Y propeller blades; is either provided on Performance data provided by propeller manufacturers the basis of an isolated propeller, or at best, with approximate correction factors based on experimental data on data of 20-year old configurations. The most recent propeller efficiency computer program compiled by Hamilton Standard under NASA contract (References 8 and 9) provide only isolated propeller information. This view is understandable for a propeller manufacturer; but for the airframe manufacturer, an important gap remains.

Current information is needed to provide a basis to determine accurate drag levels from flight-test data. The drag determined from flight can be only as accurate as the installed power basis. Improvement over the presently available data would also provide an improvement in accurately estimating installed thrust and drag and the resulting aircraft performance. Finally, improvement would provide a rational n accurate trade etween net pro eller thrust and body drag irst, a current body of empirical data i s

needed which covers the important parameters -- on current configurations with

asymmetric shapes. Second, a mathematical model of this data with current compu- tationaf fluid mechanism techniques is needed to provide a way to easily generalize data for specific cases.

Test programs could be conducted in a large-scale wind tunnel where thrust and drag can be accurately separated, As an alternate, flight test programs could also be used with special engine-propeller installations so that independent deter- mination of installed thrust can be made.

Otie such program has been proposed utilizing a separately driven propeller in the nose of a twin-engine airplane. With different nose body shapes and separately determined thrust, propeller-body interference effects could be determined. Perhaps a by-product of a flight test pro- gram would be a practical thrust-meter. This may be a l i t t l e wishful thinking, but these recommendations for targe-scale wind-tunnel tests and flight tests were proven practical by the testing performed many years ago on the obsolete configurations. So it is f e l t that these can be improved upon today, and I hope that such.a program w i l l be seriously considered.

References

1 . McHugh, J . G . 1 and Derring, E .H ., "The Effects of Nacelle-Propeller

Diameter Ratio on Body Interference and on Propeller and Cooling Characteristics," NACA TR 680, April 1939.

2. Glauert, H., "Airplane Propellers, Vol. IV of 'Aerodynamic Theory, div.

L ' , " W.F. Durand, ed.

3 . Anon., "Hamilton Standard Method of Propeller Performance Calculation, " (Black Book). Hamilton Standard, 1941.

4. Anon., "General k e d Method of Propeller Performance Estimation" (Red Book), Hamilton Standard PDB 1601, Revision A, June 1963.

5. Anon., "S.B.A.C. Standard Method of Propeller Performance Estimation," Society of British Aircraft Constructors, Ltd, 6.

Fage, A., Lock, C.N.N., Bateman, ?H., and Williams, D.H., "Experi- ments with a Family of Airscrews Including Effect of Tractor and Pusher Bodies; Part I1 Experiments on Airscrews with Tractor and Pusher Bodies, " British A.R.C. R & M No. 830, November 1922.

7 . Bateman, H e , Townsend, H.C.H., and Kirkup, T.A., "Experiments with a Family of Airscrews, Including Effect of Tractor and Pusher Bodies; Part I V On the Effect of Placing an Airscrew in Various Positions within the Nose of a Streamline Body," British A.R.C. R & M No. 1030, February 1926.

1 04 8 , vanced Genera 9 , nced General Av iat ion 1 05 eservat~on of Wing eading Edge Suction at the Plane o f Abstract ing edge near the Most fuselage geometries cover a port ion plane of symmetry, and it seems reasonable to expect that a large fraction of the leading edge suction which would be developed by the covered wing at high angles of attack is not developed on the fuselage. This i s one of the reasons that the Oswald span efficiency factor for the wing body combination fails to approach the value predicted by lifting line theory for the isolated wing. Some traditional and recent I iterature on wing-body interference is discussed and high Reynolds number data on wing-body-nacelle drag are reviewed. An exposed central leading edge geometry has been developed for a sailplane configuration. Low Reynolds number tests have not validated the design concept.

Figure 1 illustrates the significance of leading edge suction in reducing wing drag a t high angles of attack. The sketch on the upper l e f t hand s i d e of the figure gives the airload, normal to the chord, of Q symmetrical airfoil at a moderate angle of attack. The sketch on the lower left hand side compares the experimental varia-

in the absence of leading edge suction -- the

tion of drag that would be observed skin friction drag plus the load normal to the chord times the angle of attack in radians. It is seen that leading edge suction.-- the chordwise component of the

negative pressures acting on the wing leading edge -- reduces the variation of

drag coefficient with lift coefficient to a very low level for -1 < c R + 1; in

particular, for cR = -10.4, the leading edge suction i s about 5 times the skin friction drag component.

The right side of Figure 1 raises the question, for the case of a finite span wing body combination, what happens to the leading edge suction that would have been developed by the wing now covered by the fuselage, the most intense leading edge load developed anywhere on the wing .

J Lennertz('1 attempted to answer the related matter of I ift carry-over for the case of o lifting line interrupted by a cylindrical fuselage having circular cross sections and aligned with the direction of flight in 1927. ,He showed that the circu- lation at the wing root i s "mirrored" on the convex fuselage sides so that the

F 07 Preceding Page Blank

ated wing) in the fuselage concentrated longitudinal e lift carry over in th fting line, and near~y,djsappea~ in one fuselage diameter either in the up or downstream directions. The question of the leading edge suction in the chord direct ion (assuming symmetrical wing airfoil sect ions) is unresolved because the "infinitely'' long fuselage i s at zero angle of attack.

The problem of calculating the lift distribution for a wing om ion i s now routinely solved with high speed digital computers, using any of several

finite element - discrete singularity techniques for representing the flow field

around a wing body combination in such a way as to align the flow with the body surface at many control points and to satisfy the Kutta-Joukowski condition at the wing trailing edge. Such a finite element calculation has been carried out by H. Korner(*) in a form which permits direct comparison with Lennertz's analytic result. Figure 2 , taken from his paper, compares the lift distribution on wing body combinations with finite wing angle of attack, and fuselage angles of attack equal to wing angle of attack, or zero (the Lennertz case) with isolaied wing lift.

It i s seen that the zero fuselage angle of attack case agrees with Lennertz's result (Figure 1) and that the fuselage angle of attack equal to the wing angle of attack case is very similar. Unfortunately, finite element representations of the flow field near wing body combinations do not lend themselves to drag calculation, and so we do not know whether Lennertz's analytically derived relation for the reduction of Oswald's span efficiency factor for the case of zero fuselage angle and varying wing angle holds approximately for the fuselage angle equal to the wing angle or not.

we examine, in Figure 3, some model build up data in which an Next unusually large portion of the wing was covered by the fuselage and two nacelles.

The experimental drag coefficients (containing the usual corrections for attachment drag, alignment drag, and wall restraint) are plotted versus the square of the lift coefficient to aid in fitting parabolic polars to the data.

The resulting value of e, the OswaId span efficiency factor, are plotted on Figure 4 as a function of the fraction of the wing span covered, together with Lennertz's relation. It is seen that the addition of the fuselage to the wing, o r to the wing nacelle combination, causes an appreciable drop in e about twice the size ddition of the nacelles to the wing does not cause such a large charge, appar~nt~y because the wing ~ e a d i ~ e suction was the leading edges of the relatively short nacelles (there was no flow through the nacelles, which were provided with well rounded, cylindrical leading edges), but that the suction developed by the portion of the wing covered by the fuselage was not transferred to the fuselage nose, which is more than one local chord upstream. It is also difficult to justify the low Oswald span efficiency 0.804 for the isolated wing; it may be related to side edge vortex formation due to flow around the streamwise wing tip. Figure 4 also includes a point obtained in a full scale test of the Twin Commanche airplane, showing that the loss of Oswald span efficiency factor seen with XP-87 wind tunnel model i s not unusual.

Figure 5 shows an innovative geometry for a sailplane, intended to preserve leading edge suction at the plane of symmetry by the simple expedient of exposing the leading edge from 20% chord on the lower surface to 20% chord on the upper surface.

Additionally the tail assembly and the pilot's pod are attached to the w i n g by The object i s to obtain an Oswald structures of minimum aerodynamic interference.

1 and to minimize stalling at the wing-tail boom span efficiency factor approaching junction s o that CL 3/'/CD can be maximized in circuling flight to improve rates of climb in weak thermals (3).

Figure 6 presents the results of an abortive wind tunnel program to verify this design feature. A wing body combination was tested as a mid wing (leading edge covered) and as a pylon wing arrangement a t a low Reynolds number in a quiet tunnel where laminar separation might be expected to be an important aspect of the flow.

The endplating of the wing was poor a t the top of the tunnel, where air loads on the wing forced open the wing tip-ceil ing gap by increasing amounts with increasing lift, The pylon arrangement had more skin friction (the pylon plus the exposed wing area) and it i s uncertain whether the slight trend to lower drag at the highest values of CL2 is valid or not. A better test i s contemplated in the near future.

(1 1

Lennertz, J., "Beitrag zur Theoretischen Behandlung des gegenseitigen Einflusses von Tragflache und Rumpf" (Contribution to the Treatment of the Opposing Effects of Wing and Body Abhandlungen aus der Technische Hochshule Aachen Heft. 8, 1928.

Korner, H ., "Theoretische Parameter Untersuchungen an Flugel-Rumpf-

(2) Kombinationen," (Theoretical Parameter Studies on Wing Bod Combinations)

Deutsche Forschungs-und Versuchsanstalt fur Luft-und Raumfa 1: rt, Institute

fur Aerodynamic Braunschweig F B 72-63.

(3) Larrabee, E . E . , "Lateral Control and Saiplane Design Considerations to Optimize Altitude Gain While Thermalling," AIAA Paper 74-1004.

1 09 rc hl alytical tho^ in Fluid chanics Related to Drag Prediction Virginia Polytechnic Institute and State University Introduction As a counterpart to the numerical prediction methods w thought it would be useful to describe some recent theoretical work of a purely analytical nature which promises to provide engineering predictions for the important drag-related phenomena of flow in the stall regime. This analytical work deals with rigorous asymptotic studies of the complete Navier-Stokes equations that govern the where boundary layer viscous flow around any aerodynamic body under conditions separation takes place from the body surface (see for example the summary given in a recent NATO-AGARD conference proceedings ) .

Asymptotic Analysis Approach In a nutshell asymptotic analysis (sometimes also called the "method of matched asymptotic expansions" after Van Dyke or *he "multiple-deck" approach after Stewartson ) consists of a detailed local study of the Navier-Stokes ("N-SI')

-

equations behavior in the l i m i t of large Reynolds number which delineates the basic layered substructure of the flow in the presence o f adverse pressure gradients leading to separation, including viscous-inviscid interaction effects, As such, this approach is the logical next step in Prandtl's original notion o f the boundary layer theory as an asymptotic approximation for large Reynolds numbers.

The results establish how the flow leading to separation and beyond can be decomposed into a set of matched multiple decks, each of which are governed by equations appreciably simpler than the full N-S equations and hence far easier (and cheaper) to solve.

Although these mathematical studies are s t i l l rather esoteric compared to the needs of the practicing general aircraft aerodynamicist? they yield valuable insight to the "fine grain" structure of the flow and i t s basic scaling parameters (which is very useful in numerical work), and also show how to eliminate the notorious separation point singularity of classical boundary layer theory by inclusion of viscous-inviscid interaction .4 Moreover, the results of asymptotic analysis can be used to devise simplified but accurate approximate engineering solutions which are applicable over a wide range of practical flow conditions. An example of this for

Preceding Page Blank

igure 1, where the approximate triple-deck viscous- model of separation proposed by Inger i s i l l u s d schematically.

o ~ t l ~ n e ~ in Figures 2 - 4, the resulting analytical description of the flow is

relatively simple and yields closed form results for the skin friction distribution

along the surface includ h g a very useful separation point - location criterion

(Figure 4A). The turbulent flow counterpart of this layered approach can also be

ity model) , again

derived (assuming, of course, some appropriate turbulent eddy v i resulting in a very simple engineering expression for the separation point (Figure 48).

It is gratifying indeed that these theoretical results have been found to be in a good agreement with experiment over a wide range of practical conditions.

The interaction effect on the pressure distribution due to h e boundary layer displacement thickness growth 8 *(X) can also be accounted for in the contexf of this layered approach by means of a source-distribution simulation5; typical results showing the resultant smoothing out of the originally imposed non-interact ing pressure are shown in Figure 5.

Concluding Remarks With proper engineering adaptation, asymptotic analysis studies offer some valuable new methods for understanding and predicting the underlying fluid mechanics of drag (and lift) in the separation and stall regimes. Moreover, research i s currently in progress on extending these methods to include suction or blowing through the surface (i .e., boundary layer control) and the effects of Mach number (cornpressibility), including the presence of transonic shock - boundary layer i n t e r a ~ t i o n , ~ and finally to the case n f three dimensional flows. .

References 1 . Flow Separation, NATO-AGARD Conference Proceedings CP-168, May 1975.

2. Van Dyke, M., Perturbation Theory in Fluid Mechanics, Academic Press, N. Y., 1964.

3. Stewartson, K. and P,G . Williams, "Self-induced Separation,'' - Proc.

Royal SOC. A., - 312, 1969, pp. 181-206.

4. Catherall, D. and K.W. Mangler, "The Integration of the Two-Dimensional Laminar Boundary Layer Egnations Past the Point of Vanishing Skin Friction," Journal of Fluid Mechanics, - 26, 1966, pp. 163-182.

Inger, G.R., "Subsonic Laminar Boundary Layer Separation and Reattach- 5 .

ment with Viscous-Inviscid Interaction," A Y A Paper 74-582, 7th Fluid and Plasma Dynamics Conf., Palo Alto, Calif., June 1974.

6, s of Compressib and S U Aerosp sburg, V ~ r g ~ n i a , May 1975, 7 . n a ~ y t ~ c a ~ Study of Transonic No " AIAA Paper 75-853, 8th Fluid and Plasma Dynamics Conf . , Hartford, Conn , , June 1975.

= *

CY-

' W >

L a m 0 ; z x I X

I f

w .- E?

€ -r c

;;3

P

-4

s

$ '

II

If k, c

.-

c L n e , v) .

Tf he Economic Impact In General Aviation Ron D, Neal Gates Learjet Corporation Introduction ~ Historically, one of the major goals of the aircraft designer provide improved performance and it has also been recognized that one of the most significant controlling factors for achieving this goat has been the basic drag of the vehicle. As a result of the energy crisis, there has been even more current emphasis placed on the potential fuel conservation that might be derived through the incorpo- rat ion of various advanced technology concepts including improvements in aero- dynamic drag.

An example of the aircraft fuel saving benefits being considered was recently the House Subcommittee on Aviation and Transporation given during testimony before NASA officials indicated that aircraft fuel savings of up to 50% might be when achievable beyond 1985, with 5% to 10% fuel savings possible in the next few years. NASA indicated that these fuel savings would come about through 9eehnical modifications, advances in aerodynamics, structures and controls combined into a new highly efficient wing, and new materials to reduce aircraft weight .I’ The pro- jected 40% to 50% fuel savings would become available through development and integration of “optimum aircraft systems. ‘I Additional comments by other NASA officials have indicated that they place fuel saving technologies into three time levels, namely,

Near term - fuel consumption to be reduced 35% of that of current *

wide-body transports - to be achieved by 1985 - through incorporation

of supercirtical aerodynamics, composite materials, advanced pro- pulsion, advanced avionics, and active controls;

*

Far term - fuel consumption to be reduced to 55% of that of current

wide-body transports - beyond 1985 - through various boundary layer

flow control concepts; and

*

Unconventional design concepts - goals yet to be defined.

NASA is not alone in their pursuit of fuel savings for all o f the major manu- facturers are also evaluating the problem as it relates to their present and future aircraft development programs. McDonnelI-Douglas studies of a stretch DC-10 have e but there is a potentia shown that "drag reduct;ons of 3.95% are attaina e (. e ~ ~ p r o v e m ~ n t of 11.2% if a! the theoret~caldrag reduct~on could actually be gained in practice. " The ane i s reported to offer a 14% imp~ov@ment i n fuel burned per seat mile over the basis 727-200 version and for this the airlines would only pay about $2 million more per airplane. For an L-1011, Lockheed has estimated that in order to achieve a 20% improvement in direct operating costs it would require a combined 10% improvement in efficiency through aerodynamics, structures, and propulsion.

However, no matter how desirable improved fuel consumption may be, when

one considers the question of "drag reduction" and its economic impact - be it for a

transport or general avaiation airplane - it i s necessary to evaluate two factors: (1) the improvement in fuel flow (and thus lower direct operating costs) due to the drag reduction and (2) the cost associated with the incorporation of the drag reduction technology.

Before undertaking a discussion on the economic impact of drag on general aviation airplanes, it seems that a necessary first step would be to define the types of aircraft to be included i n such a study. A fairly standard definition o f general aviation is that it includes all civil aircraft except those aircraft operated in the air carrier system. The activities of this segment of aviation then range from pleasure flying by an individual pilot to the professional corporate operation of a fleet of business aircraft. The type of aircraft found in general aviation can then include the amateur built airplane, the antique, a former WWII fighter, and a business jet. Thus it is that general aviation embraces a diverse range of equipment having a multitude o f mission requirements.

Since the general aviation field includes an assortment and variety of aircraft, l e t us then determine what segment of general aviation aircraft most needs some thoughts and comments relating to the interdependence of drag and economics.

General Aviation Fuel Consumption The general aviation population may be identified, and has been identified by the FAA, in the following manner: single engine (piston), multi-engine (piston), turbine, rotorcraft (piston and turbine).

These vehicles represent a fleet of some 151,000 flying machines. In terms of flying hours in 1974, general aviation airplanes flew a total of 30,400,000 hours.

When turbine powered rotorcraft are included with the fixed wing turbine aircraft turbine powered vehicles becomes 2,640,000 hese turb~ne powered hours then represent a rai a v ~ a t ~ o n flying, In 1974, the total jet fuel and aviation gasoline consumed by the United States domestic civil aviation fleet (including general aviation) was 9,064,000,000 gallons, with the general aviation portion of this total amounting to 800,000,000 gal Ions.

From these data it is seen that general aviation consumes only 8.8% of the total domestic aviation fuel. If military aircraft operations are added to this picture, the general aviation fuel consumption drops to only 6%. When the fuel consumption of a l l forms of transportation are considered, the total fuel used by general aviation represents just seven-tenths of 1% of this total.

As a final result of evaluating these numbers, it i s noted that one segment of

general aviation, namely the turbine powered vehicles - which represent 3.4% of

the general aviation fleet and flies about 8.7% of the general aviation hours - consumes

some 44.6% of the general aviation fuel.

A further comparison of the fuel consumption of the piston powered versus the turbine powered airplane i s offered by the example that in one hour, twenty- three single-engine Cessna Model 150 airplanes will consume the same amount of fuel as one twin-engined turbofan Cessna Citation. At the large end of the business j e t scale, ninetyy-four Model 150 airplanes i n one hour will consume the same amount of fuel as one Grumman Gulfstream 11.

These numbers then clearly indicate that if a study of the effect of drag on the economics of general aviation i s to be made, the most promising area for meaningful improvement and results is in the category of turboprop and turbojet/ turbofan powered airplanes.

Operating Costs When considering the operation costs of an airplane, there are many items of expense that must be evaluated. However, a fairly common and accepted measure of the economy of an airplane is given by i t s direct operating costs.

A recent review of the direct operating costs for existing turbojet/fan business aircraft shows that the fuel cost per hour accounts for 50% to 76% of the direct operating c a t , with the average being 63%. A comparison of the turboprop airplanes shows similar trends with fuel cost averaging 54% of the direct operating costs for these airplanes.

cost per hour is direct y related to the fuel consumption of the

a~rplane - which in turn is a function of the lane drag and ef

of the engines - and to the price of the fuel I The fuel consumption of the airplane

can be controlled by the aerodynamic design and by the selection of the engine.

However, these engineering aspects of the problem have no direct bearing on the price of the fuel.

I n the 1964 to 1970 time frame, the price for jet fuel rose from 27 cents per gallon to about 35 cents per gallon.

An added burden to fuel pricing occurred in July 1970, when the Airport/ Airways Development Act went into effect. One impact of this new law was an addition of a 7 cent per gallon fuel tax for general aviation aircraft. A review of the cost analysis for a Learjet, prepared in October 1973, shows a fuel cost of 42 cents per gallon including the 7 cent tax. The same cost analysis prepared in September 197'4, used a fuel c a t of 59 cents per gallon (tax included). By November 1974, the national average for turbine fuel was being quoted at 63 cents per gallon. In the time frame of a year (1973-1974), the price of turbine fuel (excluding the 7 cent tax) increased about 60%. I n terms of an out-of-pocket this fuel price increase from 42 to 59 cents per gallon results in a $39 per expense, hour increase in the direct operating cost of a Learjet. For the operator averaging 500 flight hours per year, this amounts to an increase i n the cost of operation of 8 1 9,500 .

Airplanes tend to f l y in terms of gallons of fuel per hour or pounds of fuel per hour. However, in the petroleum industry fuel quantities are quoted in barrels rather than gallons. Airline calculations show that for every one dollar per barrel of oil cost increase, either as a result of a direct price increase or by added tax, the 2.4 cents per gallon.

price for turbine fuel increases One very simple method to reduce fuel consumption i s to reduce speed and the 55 mile per hour speed l i m i t for automobiles i s a classic example of such a s o l ut ion.

As one means of fuel conservation, the airlines are also using reduced cruise speeds. However, the impact of the reduced speeds on fuel consumption depends upon the specific airplane and i t s route structure. As an example, the b e i n g 737 can reduce i t s fuel consumption by 7% on a 500 n m . trip by decreasing the cruise

Mach number frora?O.78 to 0.7 , while incurring only a 3 minute increase in

block time. In the case of a k i n g 747, a cruise speed cut-back from Mach 0.86 to 0.84 results in a 4% fuel reduction and an increase i n block time of 16 minutes hese same trends also ho d true for the small business over a 4000 n .m . , stage lengt jet, In the case of earjet, a reduction in cruise speed from 0,81 to 0.77 will result in a t o t a ~ fue duction of about 3% over a 100 stage length with an increase in trip time of only 5 minutes. A reduction from 0.81 to 0.73 yields a fuel reduction of almost 5% and an increase in trip time of 11 minutes e Based on a Learjet fleet of 500 airplanes, with each airplane averaging 500 flying hours per year, a 5% re- of about 6,500,000 duction in fuel consumption translates into a fuel savings gallons per year.

Drag Improvements A speed cut-bock offers an operational procedure for reducing fuel consump- tion. Yet, from a long term standpoint, it i s desirable to obtain a fuel savings

without imposing a speed recfiction - even if that speed only results in a matter of a

few minutes in flight time. Looking ahead to the future the real problem to be resolved i s "What realistic improvements can be anticipated for the next generation of business aircraft ?"

A recent magazine interview with Dr. Whitcomb posed the question, "What .

new designs do you see forthcoming in the near future for corporate and general aviation?" His answer was, ' I do not think new designs of a radical nature are forthcoming in the near future, but a l l aircraft manufacturers are, of course, working on improvements to their current models . I ' This same basic viewpoint i s being echoed for the large commercial air transports and this position has been summarized as follows: "Rising costs and reduced rate of technology advances indicate a long period of derivative commercial trans- port; large technologicai advances are required to justify an all-new aircraft."

From a historical standpoint, the general aviation market has not been noted in the state-of-the-art technology. The changes for introducing major changes occurring in general aviation airplanes have tended to be in the areas of improved systems and avionics, whereas the basic airframe and powerpiant remain largely unchanged over a long period of time. This type of change does not indicate that general aviation lacks growth, for on the contrary,> the general aviation industry provides a complete range of equipment designed to meet the flying needs of today.

This observation of conservative growth is not offered a s a criticism of general aviation. If the general aviation industry were to embark on a program to incorporate high technology involving structures, aerodynamics or other advanced state-of-the-art concepts into % i s type of airplane they could certainly achieve this ss supported by military o r other government funding, the ent costs of such efforts would consideration to e faced i s that the res offer a significant improvement over the more conventional and proven co Critics of general aviation technology are a l l too ready to point out that while the airlines have grown in speed and capability through the years the general aviation airplane has remained stagnant. As proof for this premise th airlines that can be traced from the single-engine airlines of the 2 0 ' s through the modern twin-engine DC-3, the introduction of the jet powered Comet, the four- engined b e i n g 707, the new wide body transport and the supersonic transport.

However, when we examine the general aviation airplane, we also can find significant progress. The "small airplane" has developed all the way from the Wright airplane of 1903 to the high performance business j e t o f today. Thus, to claim that general aviation has not grown requires that one totally ignore and misunderstand the scope and magnitude of the general aviation market.

The real reason that the so-called " I ight aircraft" has not experienced a significant change in performance with the passage of time i s simply due to the fact that the basic laws of aerodynamics are not time dependent. Thus, it is in the real world, that an airplane of a given size, weight, and horsepower, built i n the 1970's or 80'5, will have comparable performance to a similar airplane built in the 1930's.

An excellent example of the evolution of an aircraft i s seen i n the Beech Model 35, better known as the Bonanza. This airplane made i t s first flight in December 1945. In the thirty years since i t s introduction, the Model 35 has experienced a continued history of product improvement and yet the basic airframe design, fabrication techniques and powerplant remain unchanged.

"Exceptions to the rule" do occur in a l l fields and general aviation has seen i t s share of innovative ideas. Within recent history the W indecker "Eagle" offered the promise of increased aerodynamic efficiency plus the forecast of manufacturing economy which would result in lower selling prices. Advertisements for this fiber- glass airplane clearly stated that the Eagle represented "the greatest single advance in general aviation since the advent of the all-metal airframe .I1 Yet in spite of these technical advantages, this airplane failed to achieve successful production and market status The twin-engine, two passenger "Derringer" also represented a step forward and advertisements of 1968 proclaimed: "The Derringer represents a completely advanced concept in light aircraft construction, with the same fine attention to details as found in a mil ar jet. It% the only light twin made using ch stretc~-formed, flush-riveted skins on wings and lage. All exterio are a e r o ~ n a m ~ c a ~ ~ y smooth and c ean for optimum efficiency." As with the Eagle, the Derringer failed to develop into a commercial product.

The Learjet also offers an excellent example of the continued development of an airplane. The original Learjet Model 23 made i t s first flight i n October 1963, and the delivery of the 500th airplane in April '1975 finds us with a five air product I ine. The latest addition to the Learjet family includes the Model 35/36.

The Model 35/36 i s powered by turbofan engines which offer fuel savings 30-35% over the twrbojet powered Model 25 airplane. The development of this of capability required design, development, certification and production effort covering five years. The development cost for the program was about $7 million and the airplane selling price is about $360,000 more than the Model 25.

In terms of general aviation airplanes the recent AIAA fuel workshop has provided some comments on the potential of fuel saving by means of a reduction in the

drag of "protuberances ." This workshop suggested that a full-scale drag clean-up

study of several representative general aviation aircraft be undertaken as a means of assessing the magnitude of improvement possible.

While on the subject of "roughness drag,'' I would like to offer a comment.

I find it difficult to think of a more useless effort than a study on the effects of a drag clean-up program for general aviation airplanes. It does not require a trained aero- dynamics engineer to produce a l i s t of items that, if removed or eliminated from a specific airplane, would result in some drag reduction. The real problem in a drag clean-up effort i s not an aerodynamics problem, but rather the problem i s one of how to design, manufacture and then sell at a realistic price the so-called aerodynamic improvements that have been conceived. If this area i s to be investigated, our efforts should not be spent on detailed performance improvements that might result from drag clean-up, but rather our time and monies should be spent developing economical methods of fabrication that can accommodate some of these aerodynamic changes.

In terms of an aerodynamic clean-up program, one of the f i r s t items to be considered for removai from the airframe are the antennas. As an example, we can

look at the business jet - an airplane that can fly at Mach 0.81 at altitudes of

45,000 feet - surely an airplane that would have no external protuberances to

blemish i t s high speed contours. Yet, a review of the avionics installations for this airplane shows that for any individual airplane a total of spme 13 different external antennas could be installed.

n antenna drag analy is on the Learjet has shown that if a l l of these 13 different antennas were to be ush mounted, the drag reduction w y about ?% of the tota cruise drag. As a resu t of this study, it was concluded that any flush-mounting program should encompass a l l of the antennas because the individual drag contribution of any one antenna installation is so small as to be negligible. It should also be noted that the one percent reduction in drag, due to flush mounting a l l of the antennas, would be difficult to detect i n engineering flight test since this level is within our 2 2% data scatter for cruise drag measurements.

An added consideration, to this antenna drag question, i s that for today's naviga- tion and communication duipment it i s doubtful that a l l of the antennas could be flush mounted. Thus, the actual antenna drag reduction to be realized would be somewhat less than the ideal one percent goal.

During one of the development programs on the Learjet, an attempt was made to flush mount one of the VHF antennas. The actual hardware installation of the antenna did not present a problem, however, the fact that the antenna failed to function for certain station/airplane orientations was found to be objectionable. The other factor of concern was that changing from an external antenna to a flush mount antenna involved a price change from $50.00 to about $1,000.00.

Antennas are, of course, only one source of drag in the category that may be identified as "roughness drag. 'I Included in roughness drag calculations are such irregularities as manufacturing gaps, steps, surface waves, protuberances, various air inlets and outlets, pitot probes, angle-of-attack vanes, drain 1 ines, vortex generators, and all other such items. Individually, these items usually do not produce enough drag to even be measurable from flight tests, yet taken as a sum, these items do constitute a portion of the total. Based on a drag analysis of the Learjet, it is estimated that the total roughness drag accounts for about 5% of the total cruise drag.

From an ideal standpoint it would appear to be desirable to eliminate the "roughness drag. 'I However, consideration of the engineering manhours required for the task plus the fundamental question of how manufacturing would cope with these requirements may very well lead one to conclude that "roughness drag" will remain with us for the next several years.

The supercritical wing certainly offers the opportunity for improved per- formance in tommrow's business jet aircraft. One possibility, of course, is to retrofit a supercritical wing onto an existing airplane, Yet the installation of a wing change only may not offer an economic profitable plan when the projected performance gains are weighted against the time schedule and development cost associated with this type of program.

As a specific example, in order to build two prototype airplanes with the normal development program and FAR 25 certifica- wings, conduct uire some three years and a total cost of about $8,5 m i 1 ion. In terms of airplane cost, this improvement would increase the price of the airplane about $1 50,000.

In actual fact, in order for a major change to be incorporated into a given it must offer a "significant" improvement over existing airplanes.

airplane, Conci us ions To then offer a summary, the turbine powered vehicles (fixed wing and rotorcraft) including turboprops, turbojets, and turbofans comprise a very small the general aviation fleet, yet these vehicles consume almost 45% of the segment of general aviation fuel. I n terms of general aviation fuel savings, the turbine powered airplanes offer the greatest opportunity for productive gains.

It i s possible to achieve small drag reductions through aerodynamic clean-up programs, but the improvements are usually minor relative to the engineering and i s probably on the development costs. The drag improvement from such programs order of 1 to 5%.

Improvements in airplane drag are possible within the next 5 to 10 years, but these improvements will occur on "new" models and the ir effects will be in the 5 to 10% range.

Major improvements i n fuel ccmsumption oveplexisting turbofan airplanes are real istic for 1985 and beyond, but these changes will be in the 15 to 25% range and w i l l be the combined result of improved aerodynamics - plus additional improvements from more advanced turbofan engines.

And I would hope that t h i s workshop will Serve as a springboard for the cooperation and research needed to achieve these goals in the years ahead.

5. ERS O F SESSlO Methods for Reducing Wing Drag and Wing-Nacelle Interference 5 .

T. C. Kelly, NASA Langley Research Center 5.2 Drag Reduction through Higher Wing Loading D. L. Kohlrnan, University of Kansas 5.3 Use of a Pitot Static Probe for Determining Wing Section Drag in Flight

L, C. Montoya, P. S. Bikle and E. Saltzman, NASA Flight

Research Center 5.4 Flight Test Results with an Ogee Wing T i p J. VogeI, Beech Aircraft Corporation 5.5 Wing-Tip Vanes as Vortex Attenuation and Induced Drag Reduction Devices W. H. Wentz, Jr., Wichita State University Wing Tip Vortex Drag 5.6 V. U. Muirhead, University of Kansas

Preceding Page Blank

1 35 homos C .) Kelly NASA Langley Research Center Intraduct ion This paper summarizes the results which are underway within the Transonic areas of research which w i l l be discussed are: (1) The development of both super- 1 to 3); critical and subcritical families of airfoils (see, for example, references (2) The development and application of vortex diffusers (or, more popularly, winglets) 4); and (3) The application of supercritical wings to reduce induced drag (reference to executive-type aircraft and the reduction of severe wing-pylon-nacel le interference problems which were identified during these investigations (reference 5 ) .

It should be noted that t h i s is not a summary of the total NASA-Langley effort devoted to &ag reduction, but rather a discussion of several selected areas which would be of interest.

Work at Langley i s continuing in several additional areas including the reduction of turbulent skin friction through the use of compl iant surfaces, reductions in so-called ''crud" drag through the application of various surface coverings, re- ductions in induced drag using over the wing blowing (see reference 6), tip-mounted engines (reference 7), favorable power interference effects, and the userof thick supercritical wings to achieve higher aspect ratios. Finally, the possibility of obtain- i s also under consideration.

ing practical laminar flow control Airfoil Development Supercritical airfoils - This new type transonic airfoil was developed by D r .

R.T. Whitcomb at NASA-Langley about 10 years ago. At that time, theoretical approaches for supercritical flows were nonexistent and much of the early work in developing the airfoils was intuitive in nature.

Figure 1 presents two-dimensional wind tunnel results for several ten percent thick airfoils at a section lift coefficient. of 0.7. The conventional airfoil results presented are for the NACA MA-410. Comparison of the results for this airfoil with those for the supercritical airfoils illustrates the significant gain in drag-rise Mach number achieved by use of the supercritical section. The increase in drag-rise Mach

Preceding Page Blank

y 0,l a These results also indicate the presence of drag creep which was characteristic of a 1 of the early supercritical airfoils. The which results from increases in pressure drag associated with the onset of super- critical flows on the airfoil upper surface, was noted in the early flight test results for the F8 and T-2C research airplanes which employed supercritical wings having these early-type airfoils.

Much of the recent work at Langley has been devoted to the elimination of this undesirable drag creep, and the solid curve of Figure 1 shows the result of these efforts. Refinements to the airfoil were involved primarily with changes which resulted in a more favorable flow recompression over the forward upper surface and the elimination of a region of flow overexpansion near the three-quarter chord location also on the upper surface. A slight loss in force-break or drag-divergence is noted (about 0.01) as a result of slightly increased wave losses at the Mach number higher Mach numbers, but this compromise is f e l t to be of little consequence relative to the gains achieved in eliminating drag creep. It should be noted that, unlike the early work, the shaping changes used in the design of the recent airfoil, were guided by the use of the analytical program developed recently by Korn and others (reference 8) in achieving desired pressure distributions for the various cases.

Figure 2 indicates the status o f the current supercritical airfoil development program. As i s evident, the main effort i s ambitious, and covers a broad range of design l i f t coefficients for airfoils ranging in thickness ratio from 6 percent to 22 percent. Solid symbols on the figure indicate airfoils which have been designed and tested. Open symbols represent airfoils which are currently under design using the analytical program noted earlier, and the crossed symbols indicate airfoils which are considered to have important applications and which are planned for design in the future.

Subcritical airfoils - The Langley work on supercritical airfoils led to a renewed

interest on the part of NASA in developing airfoils designed for subcritical speeds.

The first of these, designated the G A M - I , was a 17-percent airfoil designed specifically for the single-engine climb requirements for I ight twin-engine aircraft.

initial wind-tunnel results for this airfoil encopraged the design of several other airfoils for use at subcritical speeds.

Figure 3 presents a comparison of two-dimensional results obtained recently for the 13-percent thick GA(V9-2 airfoil and the older NACA 651 -213 airfoil.

Results for both airfoils were obtained in the Langley, low turbulence pressure tunnel using the narrow fixed transition strip technique. Assuming section lift coefficients and 1 . O for the cruise and climb cases, respective y, reductions in section drag coefficient on the order of 3 percent and 17 percent, respectively, are achieved for the newer airfoil a Significant gains at the higher lift coefficients are indicated.

4 corresponds to the earlier supercritical family figure and indicates Figure the current status of the subcritical airfoil family development. All of the airfoils developed thus far have been designed for a lift coefficient of 0.4, having thickness 13, 17, and 21 percent. An additional airfoil in this grou ratios of w i t h a thickness ratio of 9 percent. Also in design are two 17-percent airfoils having design lift coefficients of 0.2 and 0.7. Planned for future development are two 13-percent airfoils with design lift coefficients of 0 and 1 .2.

Vortex Diffusers A sizeable effort i s currently underway at NASA-LRC directed toward the reduction of induced drag. One phase of this effort, which has received wide notice, involves the use of vortex diffusers, or winglets as they are more popularly called, mounted at the wing tips. Unlike end plates, the vortex diffusers are designed w i t h the same careful attention to local flow conditions as would be utilized in the design of the wing itself. The placement of the vortex diffuser within the rotational flow field of the wing tip results in forward inclinations of the lift (or side force) vectors for the vortex diffusers, producing a thrust component which increases w i t h increasing lift. The action i s analogous to the force which propels sailboats, of course.

Figure 5 shows the geometric characteristics of the semispan model used for the vortex diffuser development. The wing planform represents an early version of a wide-body transport configuration. The two tip configurations tested, shown i n the right s i d e of the figure, represent the basic tip configuration and the vortex diffuser configuration which was tested assuming the "soft tip" port ion of the basic w ing panel could be remwed. This resulted in a reduction in the basic panel span of about 2.5 percent. The vortex diffuser geometry i s shown on the left s i d e of the figure.

The upper vortex diffuser span i s about equal to the basic wing tip chord; the leading edge sweep is equal to the wing leading edge sweep and the vortex diffuser root chord extends over the aft 60 percent of the wing tip chord. This particular arrangement was selected in order to avoid superimposing the high local velocities occurring on the wing upper surface and on the vortex diffuser upper surface, which faces inboard.

This upper diffuser is canted outward about 18" from the vertical a The lower diffuser i s reduced considerably in span to provide ground clearance, extends over the forward 40 percent of the wing tip chord, and is canted outboard about 3 6 ' . The upper and lower vortex diffusers were separated to avoid mutual interference effects mental drag results associated w i t h the vortex diffusers as a function of lift coefficient for three Mach numbers, In is defined as the drag coefficient for diffusers on minus the drag coefficient for diffusers off so that negative values of this parameter represent gains or, thrust.

Near zero lift, a net penalty results as would be expected. At a lift coefficient of about 0.26 the diffusers are carried with no penalty, and above this I ift coefficient , favorable effects are obtained which increase with increasing I ift .

For the case presented, the cruise lift coefficient i s about 0.53 at a Mach number of 0.80, resulting in a gain of about 15 drag counts ( C , , = 0.0015). In full-scale terms this would represent.an increase in L/D of about 5 percent, which, of course, is significant .

Figure 7 indicates the effect of the vortex diffusers on the wing pitching- moment and root-bending-moment coefficients, These results are for the cruise Mach number of 0.80, and show relatively small effects of the diffusers, for example, a two percent increase in root bending moment at the highest lift coefficient tested.

Also, early wind tunnel flutter tests have indicated relatively small reductions in flutter dynamic pressure resulting from addition of the diffusers. It was also concluded that these effects were associated with structural characteristics rather than any unsteady aerodynamic interaction Figure 8 illustrates one proposed application of vortex diffusers. NASA is currently involved in a joint program with the USAF to determine the possibility of adding vortex diffusers to both the C-141 and the KC-135. Tests df both configurations are scheduled for this fall in the 8-foot Transonic Pressure Tunnel. This model photo- graph depicts vortex diffusers installed on the C-141.

Wing-Nacelle Interference In an effort to provide access to supercritical wing technofogy for the general aviation manufacturers, NASA has entered into several cooperative endeavor agreements with members of the industry whereby NASA provides expertise in the areas of aerodynamic design and application of supercritical wings to this class of aircraft and also provides limited wind tunnel testing for configurations which are under design. The remainder of this discussion w i l l present some results obtained recently which relate to the problem of wing-nacelle interference which occurs at h@h subsonic speeds and which i s characteristic of configurations where fuselage-mounted engines overhang the wing rearward upper surface. A representative configuration is shown in Figure 9. This is a photograph of a one-ninth scale model of an 1 40 executive-type aircraft under test in the Langley eight-foot transonic pressure ar investigation was conducted t o determine the ae ity of replacing the original wing w i t h one having supercritical sections of increased thickness, the thickness ratio being increased on the average from about .09 to .12. Because the wing was intended to be retrofitted, other configuration changes were to be kept to a minimum, therefore most of the tailoring or "tuning" itself although some cha changes were associated with the wing were also made.

Figure 10 illustrates the severity of the interference problem which was found to exist between the wing and.the engine-pylon arrangement for the modified con- figuration. These results are for a lift coefficient of 0.25 and are presented as an incremental drag coefficient versus Mach number, where M = 0.60 i s used as a reference drag level for each configuration. It should be emphasized that a l l of the results presented are for the configuration with a supercritical wing. No comparisons with the basic wing.

are presented The data for the initial supercritical wing configuration, shown by the solid curve, indicate noticeable drag creep and an early drag rise. Reasons for these effects will be discussed later. Removal of the engines and pylons resulted in sig- nificant reductions in drag-creep, and the force-break o r drag-rise Mach number was increased by about 0.02.to 0.03 as indicated by the dashed curve. Modifications to the wing root section, the addition of a wing glove and some reshaping of the pylon provided the results indicated by the long-short dashed curve. As, can be seen, the drag creep was reduced considerably (about 75 percent at a Mach number of 0 . 8 0 ) .

Figure 11 presents wing upper-surface oil flow photographs for a Mach number of 0.825 and a lift coefficient of about 0.35. The photograph on the left of the figure is for the initial configuration w i t h nacelles and pylons and shows dramatically, the effect o f the nacefle and pylon presence in forcing the uppers urface shock wave forward on the inboard wing region. The presence of a second wave which originates in the channel formed by the upper surface of the wing and the nacelle-pylon combina- tion is also apparent. Examination of local pressure distributions for this case indicates the second wave and the adverse pressure gradients 'associated with it caused extensive separation in the "channel" region. The center photograph, for the case of the nacelles and pylons removed, shows what could be termed an expected supercritical wave location for this thickness ratio on the upper surface with little or no separated flow in evidence.

photograph of lustrates the upper surface condition for to er co the tuned configurat ion e The main wave appears somewhat for the initial configurat~on,and no evidence of the second wave i s seen Modifica- tions to the configuration were accomplished through a number of steps. Figure 12 shows a comparison of the initial wing root and pylon lines with those for the final configuration tested. Most of the changes noted were made to eliminate what was essentially a converging-diverging channel formed by the w ing-pylon-na combination. Initially, material was removed from the aft region of the wing upper surface. Then, in order to reduce the upper surface velocities entering the channel, a glove was added and the forward portion of the resulting airfoil was increased in thickness. Finally, the lower surface of the pylon was thickened in order to provide h for a relatively constant area in the channel, A l l o f the noted changes represented an application of the area rule in a local sense, since the induced velocities approached and exceeded sonic values in the region.

The results just disaussed were obtained using an executive-type aircraft model which employed turbojet engines. It might be expected, therefore, that replacement of the turbojet engines with the newer and larger turbofan engines would .

serve to aggravate the interference problem which was noted and some recent wind tunnel results indicate this to be the case.

Figure 13 presents results for an executive-type aircraft very similar to the one previously discussed. Again, results are presented in the form of an incremental drag coefficient versus Mach number; however, the reference Mach number i n this figure i s 0 . 5 0 . Because an interference problem was anticipated, provision was made on the model to allow translation o f the engine-pylon combination to several longi- tudinal locations. The results given in Figure 13 are for the proposed initial location (identified in Figure 13 as the production location), the most rearward location possible, which corresponded to a full-scale rearward shift of 18 inches and for a final "tuned" configuration with the engines aft. For this span location, and neglecting the glove, the leading edge of the nacelle moved from about the 55-percent chord location to the 70-percent location. This relochtion was made in three steps of 6 inches each, full scale, and the results indicate that further gains could be achieved by additional rearward movement of the nacelle. Obviously, however, airplane balance problems would impose some practical rearward limit to the nacelle reloca- tion. As with the earlier results, noticeable additional gains were made through so-called "tuning" which, in this instance, involved primarily changes in the inboard a irfoi I shape.

1 42 ransonic Aerodynamics Branch at NASA-~RC t ion in several areas e rirnary efforts have been involved w i t h the design of both supercritical and subcritical families of airfoils, the reduction of induced drag through the use of vortex diffusers, and the reduction of interference drag for executive-type aircraft .

The results of many of these efforts are felt to be applicable to the design of general aviation aircraft.

References 1. Whitcomb, R . T a , "Review of NASA Supercritical Airfoils," ICAS Paper No.

74-10, The Ninth Congress of the International Council of the Aeronautical Sciences , August 1 974.

2. McGhee, R . J . , and Beasley, W.D., "Low-Speed Aerodynamic Characteristics of a 17-Percent-Thick Airfoil Section Designed for General Aviation Applica- tions, NASA T N 0-7428, 1973.

McGhee, R.J., Beasley, W.D., and Somers, D.M., "Low-Speed Aero- 3.

dynamic Characteristics of a 13-Percent-Thick Airfoil Section Designed for General Aviation Applications, NASA TM X-72697.

Flechner, S.G. , and Jacobs, P.F., "The Effects of Vortex Diffusers (Wing-

4.

lets) on a Semispan Model of a Representative Wide-Body Transport," Proposed NASATN D.

5 . Bartlett, D.W., "Application of a Supercritical Wing to an Executive-Type Jet Transport Configuration," Proposed NASA TM X. (L-9939).

6. Putnam , L .W., "Exploratory Investigation at Mach Numbers from 0.40 to

O a 9 5 of the Effects of Jets Blown Over a Wing," NASA TN 007367, 1973.

7. Patterson , J D. , and Flechner, S .G., "An Exploratory Wind-Tunnel Investiga-

tion of the Wake Effect of a Panel Tip-Mounted Fan-Jet Engine on the Lift- Induced Vortex, '' NASA T N D-5729, 1970.

8. Bauer, F. , Garabedian, P . , Korn, D. , and Jameson, A,, "Supercritical

Wing Sections II," Lecture Notes i n Economics and Mathematical Systems , Springer-Verlag, c 1975.

3 \

cn

I

I

\

\

a

v)

\'\ \'

1 44 x

.-

u E \ U c u .

Q p -

0 0 0

r C

0 c3

/

/

.

*

x

. . _. .-

E i

U

0 e

1 0 - 7 0 F: Q I & cu

- . v)

v) w .

-4.

(0.5

0 - m .I "I L L I- N a? c ' r q e 1 47 0 . I ,2 . 3 . ' .4 .5 .6 Figure 6. Vortex Diffuser Effects on Drag Coefficient 1 49 .a -24 0.20 .I6 % W R .

: R

.m

.w

-.w * -

-.I 0 .I . Z .3 . .I .S ' .6 .7 Figure 7. Vortex Diffuser Effects on Pitching-Moment Coefficient and Wing Root Bending-Moment Coefficient 1 50 1 52 a i I

P

t o

X

w

L I 1 I a0 d=

v 3

ro

2 cn

- .-

m 0 2

I-

-

z

b

z

Q

a

c N c c 1 55 c n v)

w

z

W Y I- LL

’ \

ag eduction Through Higher Wing loading

vid L. Kohlman

~ n ~ v e r s i ~ of Kansas I ntroduc t i on The wing typically accounts for almost half of the wetted are production light airplanes and approximately one-third of the total zero-lift or parasite drag. Thus the wing should be a primary focal point of any attempts to reduce drag of light aircraft with the most obvious configuration change being a reduction in wing area. Other possibilities involve changes i n thickness, planform, and airfoil section.

This paper w i l l briefly discuss the effects of reducing wing area of typical light airplanes, constraints involved, and related configuration changes which may be necessary.

Constraints and Benefits The wing area of current light airplanes i s determined primarily by stall speed and/or climb performance requirements. Table I summarizes the resul ting wing loading for a representative spectrum of single-engine airplanes. The maximum I ift coefficient with full flaps, a constraint on wing size,is also listed. Note that wing loading {at maximum gross weight) ranges between about 10 and 20 psf, with most 4-place models averaging between 13 and 17. Maximum lift coefficient with full flaps Panges from 1.49 to 2.15.

Clearly i f CL can be increased, a corresponding decrease i n wing area can max be permitted with no change i n stall speed. I f total drag is not increased a t climb speed, the change in wing area will not adversely affect climb performance either and cruise drag will be reduced.

Though not related to drag, i t i s worthy of comment that the range of wing loading in Table I tends to produce a rather uncomfortable ride in turbulent air, os every Iight-plane pilot i s well aware. The only way to reduce this gust sensitivity is to increase wing loading. ' Typically, wing loading tends t o increase as performance (cruise speed) increases. This i s particularly evident in Table I1 which presents data for twin-engine aircraft. But gust response i s proportional to the ratio of calibrated cruise speed to wing loading (VJ(W/S))and thus improvements in ride due to higher wing loading are partially, if not completely, offset by higher cruise speed.

It i s also evident in T a l e II that even though wing loading i s higher than f o r eeds . Twi n-engi ne s i n g ~ e ~ e n ~ i n e aircraft, i t i s translated directly into higher stall high lift systems oduce virtually the same C le I for single-engine L l O X airplanes. Thus ere appears to be an equal potential for reduction i n wing area of single- and twin-engine aircraft by employing improved high lift systems, tiow for light aircraft i s discussed later.

to achieve higher CL

But assuming E Y a moment that improvements i n C are a , making

' m a , higher wing loading possible for a given airplane or class of airplanes, i t i s important to consider how the wing area ihould be reduced. The easiest and most tempting way i s by reducing span. N o t only does this leave the inboard wing structure, mechanisms, and wing-body junction unchanged, but i t reduces wing bending moments making possible a lighter wing. But reducing the span increases the span loading, thus reductions in parasite drag through a decrease in wing area are countered by an increase i n induced drag.

On the other hand, reducing wing area by a decrease in wing chord decreases parasite drag almost i n direct proportion to chord decrease, and i f span remains constant there is virtually no change in induced drag. From an aerodynamic point of view this i s most desirable, but i t introduces possible structural and weight problems because aspect ratio increases A i l e spar thickness and internal volume decrease if the same airfoil section i s used.

To understand the potential and the constraints of drag reduction through wing area reduction, consider the following simplified analysis.

Assuming that the parasite drag coefficient and span efficiency factor remain unchanged, the parasite drag i s directly proportional t o wing area and induced drag is inversely proportional to the square of the span. Then the wing drag at any given flight condition may be written as where Dp i s h e reference wing profile drag; Dw i s the total drag of the reference R R wing. The span and chord are denoted as b and c with a subscript R indicating reference values. For simplicity an untapered'wing is assumed.

Normalizing equation (1) with respect to the original reference wing drag, D wR ' gives , the ratio of parasite drag to total drag.

If only h e wing chord i s reduced, then the change i n total n o r m a ~ i ~ e d wing drag i s dc

dD= P -

(3) C R Thus the percent reduction in total wing drag i s equal to the percent reduction i n chord length times the original ratio of parasite to total wing drag. Clearly, the benefits of wing area reduction increase with air speed.

Consider a typical light airplane with the following characteristics: Gross weight = 2800 pounds Aspect ratio = 7.4 Wing area = 174 ft Drag coefficient of body and empennage, CD = 0.017 OBVH = 0.009 Wing parasite drag coefficient, CD O w Airplane efficiency factor, e = 0.75 Cruise altitude = 8,000 f t If only the chord i s reduced, then,as shown i n Reference 1, the resulting

normalized total airplane drag, DT, i s shown in Figure 1 . AI though substantial drag

reductions are possible, constraints are imposed by the requirement to cruise a t a reasonably low l i f t coefficient and stall margin, and to keep stall speeds’low enough for good takeoff and landing performance. Even with these constraints, however, significant reducfions in wing area, cruise drag, and gus+ response are possible for today’s general aviation fleet.

To analyze the effect of reducing span while holding chord constant, differentiate equation (2) with respect to span b. Then

For a decrease i n span to result in a net decrease i n drag the condition g > O

for b = bR must be satisfied.

This i s )rue only if

P>F 2

(5) I n other w o r d s , a reduction i n drag by reducing span can be achieved only i f parasite drag is more than double the induced drag at the flight condition in question.

While this may be satisfied during high speed cruise, i t i s rarely true during a climb,

And when P < 2/3 a reduction in span increases induced drag more than i t decreases

parasite drag. For a tapered wing, P must be even larger than the value given i n (5) to achieve drag reduction.

b The l i m i t to favorable span reduction is found by solving for the value o f which yields db dD = 0, assuming P>2/3. Again from equation (4) i t i s easily shown that

2 (1 - P)

' when (e7 = P

Equation (6), plotted in Figure 2, establishes the boundary of favorable span reduction of a constant chord wing a s a function of the reference wing parasite drag ratio, P.

Technical Devetopmen ts I t i s clear that wing area reduction can be achieved only if corresponding can be designed into light airplanes i n a practical manner.

increases i n CL max Several recent developments indicate that this i s a very real possibility.

One promising development is a new family of general aviation airfoil sections. Two members of the family, the GA(W)-1 and GA(W)-2, have been defined As shown i n Reference 2, the characteristics of these airfoils are: at this time.

-

high C L , ~ compared to conventional airfoils (see Figure 3)

- gentle stall Characteristics

- fairly thick section. The GA(W)-I i s 17% thick. This helps to

maintain spar depth with reduced chord lengths.

- very little increase i n C D a t climb lift coefficients (see Figure 4).

This combined w i t h 0 decreased w i n g area offers the potential of significant increase in single-engine climb performance of twins.

Another interesting development i s the recognition of the efficiency of spoilers for roll control on light airplanes. Among other features, spoilers permit the use of with no change full-span, or a t least increased span, flaps. This w i l l increase CL max in airfoil or flap geometry. Several light airplanes are now using this concept: the advanced technology light twin (ATLIT), a modified Seneca; the Redhawk,, a modified Seneca, a rnodific tion k i t developed by Robertson Aircraff Cessna Cardinal; the Corporation; and the nother method of increasing C i s to increase the Fowler action of hnax conventional single-slotted flaps. This can be done with very little increase i n complexity or weight. Figure 5 shows the very large values of CL (2-D)which max can be obtained with a GA(W)-l airfoil using a 30% chord single-slotted Fowler flap.

Flight Test Results Additional confirmation of the ability to increase C through both airfoil Lmax design and flap design has been demonstrated in the Redhawk and ATLIT programs.

Table 111, from Reference 3, shows maximum lift coefficients obtained on the Note that the flap covers Redhawk by using a 30% chord single-slotted Fowler flap.

only 47% of the wing span.

The ATLIT, using full-span, 30% chord single-slotted flaps, and a GA(W)-1 basic airfoil, generated the high l i f t data shown in Table IV. Clearly, significant increases in CL are possible for this class of airplane.

m a X Finally, Table V shows drag data generated during flight test of the Redhawk.

The most significant result i s that parasite drag was reduced 10.5% by reducing wing area, thickness, and span. This i s a significant reduction, and i t illustrates i n flight that a reduction i n wing area can be an effective and practical means of reducing drag.

References 1 . Roskam, Jan, "0 portunities for Progress in General Aviation Technology, I t A I M Paper No. 75-292, presented at AIAA 11 th Annual Meeting and Technical Display, Washington, D. C . , February 24-26, 1975.

McGhee, R. J., and Beasley, W. D., "Low Speed Aerodynamic Characteristics 2.

of a 17-Percent-Thick Airfoil Section Designed for General Aviation Appl i- cations, I' NASA T N D-7428, December 1 973.

Kohlman, David L . , "Flight Test Results for an Advanced Technology Light 3.

Airplane Wing, I' SAE Paper No. 740368, presented at Business Aircraft 2-5, 1974.

Meeting, Wichita, Kansas, April able I.. g" Loading and C for Typical Single-Engine rnax C

A~rcraft s - PSF max

Cessna 150 10.2 1.73 Cessna 172 13.2 2.15 Cessna 182 16.9 2.03 Cessna 210 21.7 2.01 Beech C23 16.8 1.89 Beech V35B 18.8 1.85 Grumman Tiger 17.1 1.92 20.6 Bellanca 300A 1.64 Mooney M20E 15.4 1.85 13.4 Piper PA-28-140 1.73 Piper PA-28-180 14.4 I .51 Piper PA-28-200R 15.6 1.49 Piper PA-32 19.5 1.92 Table 11. Wing Loading and C for Typical Twin-Engine Aircraft Lmax W/S- PSF cL max . Aircraft Beech Baron 25.6 1.42 Beech Duke 31.8 1.64 Beech Queen Air 28.9 1.78 30.7 2.02 Cessna 310 Cessna 402 32.2 2.02 Cessna 421 35.2 1.86 Piper Seneca I1 21.9 1.80 Piper Navaho PA-31-350 30.6 1.66 Piper Navaho PA-31 P-425 34.1 1 *93 1 62 Comparison o f Stall Speeds an Lift Coeff icienk Configuration Redhawk Cardinal Cruise 79.6 1.40 64.7 1.35 Kruger flaps only 69.8 1.82 I

Fowler flaps 10" 71.2 1.75 -

Fowler flaps 10' and Kruger flaps 62.8 2.25 Fowler flaps 40' 64.4 2.14 55.0 1.84 (300 for Cardinal) Fowler flaps 400 and Kruger flaps 56.0 2.83

Notes: 1 . Gross weight = 2500 Ib

2. Redhawk c.g. location '7.2% m.a.c. (109 in.)

3. Cardinal c.g. location 19%m.a.c. (109.3 in.)

re1 imi nary Stal I P t i

cL

max f 0 76 1.81 1 O0 66 2.40

20° 61.5 2 .n

3 0 ' 59:3 2.98 40' 59.4 2.97 Gross weight = 4200 Ib Aft c.g. location Table V. Comparison of Drag Characteristics Determined from Flight Test C S 'D w DP P Cardinal Cruise 0.0267 4.67 Full Flaps 0 .OM2 8 . 0 8 Redhawk Cruise 0.0380 4.18 Full Fowler and Kruger F 1 aps 0.0788 8.67 1.0 .8 .6 .4 .2

. 6 .7 . 8 .9 I .o

P

Figure 2. Limit of span reduction to decrease drag as a function of parasite drag ratio.

NASA G A ( W ) - I airfoil

Roughness off Roughness on NACA airfoil ,roughness off ( r e f . 5 )

---

652-415 --- 4418

2.2

E -- 653-418 ---- 23018 fftfistfitf

2.0 1.8 (C2)rnax 1 . 6 1.4 1.2

I .o

0 2 4

I O 12 i4xlO”

Figure 3. Variation of maximum section lifttoefficeint with Reynolds number for various airfoils without flaps. M = 0.15.

1 67 N A S A GA(W) - I airfoil 0 N A S A standard roughness fl NA C A standard rouc~ttncss

N A C A airfoil , NACA

standard roughness (ref.5 ) - -- -- ~ 652- 415 ....

.04 . . . .

. . .

. . .

653- 418 . . . .

. . .

. . . .

. .

. . . .

. . . .

.03

. . . .

! . . ^ . , . . . .

. . . .

.. ..- . . .

, ' . ' . . . .

. . . .

- - .02 I . . .

'd . . .

. , . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

_ . _ .

.o I

. . . .

. . . .

. . .

. . , .

. . . .

. . .

. . . .

. . . .

. . .

-

-.% -.4 0 ,4 2.0

I .2 I . 6 C t Figure 4. Comparison of section drag characteristics of NASA GA(W)-l airfoil an NACA 6!j2-415 and G3-418 airfoils. M = 0.20;

Rz6x10. d

.. ..

. ...

' : i " , : I : . : : ! . .

. .

'I Figure 5. Lift erf rmance of a 3 % cgo d. single-slotted FOWL fP,p on a GA(W"-I ~ A A .

5.3 Useofa itot Probe for ~eterm~ning Wing Section Drag in Ftight Edwin J a Sal tzman NASA Flight Research Center (A status report of development work of Lawrence C. Montoya and Paul F. Bikle) Introduction This paper presents the status of recently completed development work on the wake traverse method o f obtaining section drag at low speeds. The method of B. M.

Jones, Reference 1, has been applied to wake profile data obtained from the wing of a sailplane.

Though’ the sailplane provides a quiet and relatively vibration free environment, and a very clean and smooth high aspect ratio wing, it i s limited to a dynamic pressure range from about 6 psf to somewhat above 20 psf. These low dynamic pressures are a severe challenge for the wake traverse method.

This paper i s intended not only to rekindle interest in the wake traverse approach of defining profile drag, but it i s also intended to demonstrate techniques for increasing reliability and minimizing certain bias errors when dealing with relatively low differen- tial pressures. The thought which accompanies this paper i s that i f acceptable accuracy i n section drag can be obtained at these low dynamic pressures, even better profile drag definition could perhaps be obtained for applications on general aviation aircraft where higher pressures would prevail.

Airplane and Tesf Conditions The sailplane was a T-6 having a modified Wortmann FX-61-163 airfoil. The modification consisted of a straightening of the underside cusp region at the rear portion of the section. A sketch of the wing profile i s shown i n Figure 1 .

The surface finish of the wing was very smooth and the maximum waviness was 0.003 inches in a two-inch section of surface. At the semispan station of the about this.

wake traverse tests the waviness was less than The wake measurements were made 9.6 inches behind the wing trailing edge which corresponds to about 32 percent of the 29.9 inch local chord. Data were obtained for speeds from about 40 knots to 125 knots which provided .. chord Reynolds 6 b numbers between 10 and 3 x 10 .

Preceding Page Blank

a~rspeed was held constant during each data run and six "total-pressure" re made during h run followed by three " s t u t ~ ~ ~ p r e ~ u r e " wake traverses. This sequence of nine traverses took approximately one minute. Data been obtained for deflected flap conditions but the data to be shown herein are for zero flap deflection. The majority of the data were obtained for very smooth air conditions but some data were obtained for air that would be considered somewhat rough.

A photograph of the sailplane with the wake traverse probe installed i s shown in Figure 2. A Kiel probe can be seen mounted ahead of the drive unit package on the right wing. The dark colored tape running parallel to the wing trailing edge covers and holds down the wiring harness which connecis the drive unit puckage to the recording package inside the fuselage.

Instrumentation Probe and Drive Unit: A cioser view of the probe and drive unit, and some of the reference probes, i s provided i n Figure 3. The Kiel tube i s used as the reference for the transversing total pressure probe and the trailing boom provides a static reference for the traversing static pressure probe. The trailing boom orifices are located about 5 feet behind the wing and are calibrated against the ships static system, which i s in turn a system which has been calibrated for position error.

The traversing probe (with total and static heads) and drive unit ore shown again in Figure 4(a) and i n closer detail i n Figure 4(b) where some of the ports have been labeled. The probe traversed to about 8 inches above and below the wing trailing edge. The probe travel rate was a l i t t l e less than 3 inches per second. The hardwore which i s shown in Figure 4 weighed about 3 pounds.

A very important part o f the unit was the switching valve which permitted the same pressure transducer to measure the wake station total pressure decrement, in one mode, and the difference between wake static and trailing boom static pressure, when i n another mode. More detail w i l l be provided about this feature in following figures.

Recorder Package: The recording package consisted of a tape recorder, battery and a component box which housed a pressure transducer and two amplifiers. This package was mounted on a shelf behind the pilot's head rest as can be seen in Figure 5 .

A closer view of this hardware i s shown i n Figure 6, after a covering hatch has been removed. The weight of this package i s about 40 pounds. The switch shown in Figure 6 i s also an important element in obtaining in-flight tare readings on the transducer which records "ships q."

igure 7 w i l l be used t o show schematically how the tioned i n the previous section works. As sketched in Figure 7, roing mode." When the pilot places the switch in this mode, side of the pressure transducer element, thus freestream total pressure exists on each providing an in-flight tare reading which minimizes the bias error of this transducer.

When the switch is placed in the "q mode" (toggle to the left) there i s freestream total pressure on one side of the transducer and ships static on the other side. T course, provides a record of "ships q." Because the ships static system has been cali- brated for position error the appropriate corrections are applied and a true freestream dynamic pressure can be calculated as a function of time for correlation with the pressures recorded i n the wake and the probe position i n the wake. The airspeed indicator allows the pilot to hold "ships q" steady for a sufficient period of time to permit up to nine successive wake traverses to be made under quasi-steady state conditions.

Another switching feature involves the switching valve which was identified in Figure 4(b). I n Figure 8, a slide valve i s shown in schematic form to illustrate how this switching valve can be used to direct Kiel tube pressure (freestream total) to one side of the transducer and wake total pressure to the other side. Thus the transducer senses the total pressure defect, APT, i n the wake. When the wake probe moves beyond the wake, the transducer experiences freestream total pressure on both sides of the sensing element which thereby provides.inflight tare readings for the transducer.

This feature minimizes the bias error for this transducer.

Another mode for this switching valve i s illustrated i n Figure 9. In this case the valve arrangement provides trailing boom static pressure to one side of the sensing element and wake station static pressure on the other side. As mentioned before, the trailing boom static pressure has been calibrated against the ships static system which i n turn has been calibrated for position error. Therefore the true decrement between wake station static pressure and freestream static can be calculated. I t is this corrected decrement, AP, which appears as an adjustment i n the Jones expression for calculating section drag from wake measurements. The Jones equation follows: where

q , freesiream dynamic pressure, from ships transducer plus position error

correction ressure decrement i n wake , from probe transducer ke

P , difference between freestream and wake static pressure, from probe

ha ns ducer dy, from probe position potentiometer c, wing chord a t wake survey location

+ AP , fromq -APT

qwake wake qwake I t i s important that the same transducer, through the switching valve feature just

described, provides both A P , and L I P because thereby the parameter AP

. Twake has bias error minimization i n the same way as has been described for the AP Twake measurement, I t i s also important to note that s i x AP traverses and three A P traverses Twake are made in succession while indicated airspeed, and consequently q , is held constant.

Section drag Coefficients to be presented i n a following section are the average of six such traverses for each Reynolds number condition.

Results Profiles of A P plotted as a function of distance above and below the Twake wing trailing edge plane are presented i n Figure 10. These are obtained from six consecutive traverses through the wake over a period of about 40 seconds. The airspeed was about 44 knots which resulted in a section lift coefficient of about 1 .O. Note the low "delta pressures" with which the instrumentation must contend a t these speeds.

Some part of the apparent dispersion of the six profiles i s random scatter; however, a part of i t i s caused by small changes i n airspeed during the 40 seconds of tun time.

Therefore, a part of the apparent dispersion will be eliminated when each profile i s normalized by the appropriate mean q for thbt traverse. As mentioned before, the six normalized profiles are then averaged when calculating section drag coefficients by Jones' equatiow .

Another set of profiles is shown i n Figure 11 for an airspeed of about 42 knots which provided a lift coefficient near maximum. I n this case only, four profiles are 1 74 he profile with í angular symbols i s the f i r s t of the sequence and i s similar i n appearance to those shown i n Figure 10. The following profiles i n their order of occurrence (diamonds and squares) progress toward local stall (circles) as speed was reduced ever so slightly. Earlier flying with flow visualization indicated that this region of the wing was indeed stalling a t these speeds.

Section drag coefficients plotted as a function of chord Reynolds number are shown in Figure 12. Also shown are curves from Blasius and Schlichting representing fully laminar and turbulent flow over the upper and lower surfaces as if they were flat plates.

The circles represent three flights (each point i s the average of 6 traverses) for natural transition. The wing surface was exceptionally smooth and clean for these runs.

The squares represent flight with the boundary layer tripped 5 percent behind the leading edge for both the upper and lower surfaces. The trip material was dis- tributed'grit of 0.035 inches mean height above the skin surface.

The results obtained to date have been encouraging and suggest several

possibilities which may deserve to be included i n fotiow-on flighis . Some of these

possibilities are listed below: (a) . 400 grit sanding of the local surfaces (b) 200 grit sanding of the local surfaces (c) add waves, up to 0.008" per 2 inches (d) restore to best finish (fill, sand, rub) (e) tape over flap gap insect roughness (bugs) on leading edges (f) install wake traverse unit on several general aviation aircraft .

(9) Closing Remarks The experience reported has provided progressively improving accuracy i n defining section drag for flight at low dynamic pressures. The accuracy i s about 3 f o 4 percent for the lower dynamic pressures and somewhat better than 3 percent at the higher dynamic pressures. I t should be emphasized, however, that such accuracies were obtainable only by resorting to the detailed pr'ocedures described herein (especially the in-flight switching and tare evaluations) and by conducting a l l phases of the testing with the utmost care.

Also of great importance was the precise position error calibration which had been accomplished on the ships airspeed-altitude system prior to the wake traverse work.

er care and attention to detail, comparable or accurate section drag coefficients cowid be obtained on general aviation aircraft, ~ o n s i d e r ~ n ~ the higher dynamic pres$wr~ encountered in general aviation.

References Reference 1 . Jones, B. M., Measurement of Profile Drag by the Pitot- Traverse Method, Cambridge University Aeronautics Laboratory, Rand M.

N o . 1688, 1936.

1 7 6 I J 77 Q1 h 9) c

Y- f

(u & c

.-

C a .

I U a m

.^

‘I 83 c

t-

z

w

I

w

! ' cc c

.-

L

z

R

L

cn

.-

LL Q ! !

1 84 V Sr, 44 KNOTS r ~ ~~ 6 CONSECUTIVE WAKES VI w X u

E i

..

H n

0 1 I

,1.4 2 . 8 A P , , PSF Figure 10. Example Wake Pressure Profiles (44 kts) Vc'Y 42 KNOTS I 3 1 B z U .I

!l E

r n w n

a

x

E 2

m X I

I 1

a 1 . 4 2.8 A P , , PSP Figure 11. Example Wake Pressure Profiles (42 kts) 1 88 I

i

I

I

I

I

I I

I

I

I

I

I

I '

I

J

I

I

I

I

I

I

I

I

I

I

I

t

4 0 I 1 89 Introduction Wind tunnel tests by NASA indicate that the aerodynamic performance of a rectangular 3-D wing can be increased by changing the tip to an ogee shape. Test data obtained during the tests show substantial gains in L/D throughout the angle of attack range of interest.

In order to investigate the potential gains. in both cruise and climb performance, a Beech Baron was modified to include a pair of ogee tips on the configuration.

Estimated gains i n performance based on rectangular wing test data were somewhat optimistic. Increases in cruise speed and climb rate were predicted to be as much as 5 mph and 100 fpm, respectively. A series of quick tests were scheduled to see if the predicted gains could be realized in practice.

Ogee T i p Review few years has been in conjunction with Ogee tip research during the past rotor blade study. One of the basic problems associated with rotor blade flows is the concentrated tip separation vortex generated with rectangular tips which degenerate the quality of the flow field encountered by the following blade. The ogee tip i s designed to eliminate or diffuse the separation vortex. This is accomplished by cutting back the tip streamwise edge, starting at the leading edge as shown in Figure 1.

Wind tunnel tests by NASA to indicate that the separation vortex can be diffused w i t h the ogee tip. Figures 2, 3, and 4 show upper surface isobars for a rectangular wing and an’ogee tip section with the same wing are. (Plots taken from reference 1 .)

As noted, the separation vortex i s eliminated.

Balance data indicate that the decreased primary vortex activity leads to a substantial increase in wing L/D. Figure 5 shows the comparison between the rectangular wing and the equivalent ogee tip configuration.

The Modified Beech Baron The Beech Baron was chosen as the test bed for two reasons. First, it has a detachable tip to easily accommodate the ogee tip; and second, the higher performance allows a chance for greater absolute changes in incremental performance, thus improving the flight-test accuracy.

Preceding Page Blank

and the ogee tip is The wing areas are the ant 23012 section a1 F I ight Test Results FI ight tests included both speed-power and sawtooth climbs. Results, as noted i n Figures 8, 9, and 10, indicate that incremental changes in performance due to the addition of the ogee tips are on the order of the data scatter associated with flight test techniques used. Climb data a t density altitudes of 5500 feet and 9500 feet show a possible increase in climb at high CL values and a decrease at low CL values. The speed-power data indicate no substantial change in level-fl ight speeds.

References 1 .

"Effect of Sweep Angle on the Pressure Distributions and Effectiveness of the Ogee Tip in Diffusing A Line Vortex," J.C. Balcelak and R.F. Feller, NASA CR 32355 2.

Rorke, J.B., Moffitt, R.C., and Ward, F.J., "Wind Tunnel Simulation of . Full-Scale Vortices," Preprint #623, 28th Annual National Forum of the American He1icopter Society, Washington, D .C., May 1972.

Landgrebe, A.T., and Bellinger, E.D., "Experimental investigation of Model 3.

Variable - Geometry and Ogee Tip Rotors," United Aircraft Research

Laboratories, NASA Contract No. NAS - 10906, to be published, 1973.

Figure 1 . Ogee Tip Planform

.o I

I

Y/b

cP

.02

I

. O 3

Z W 5: w w -J i I

-. 8 - 1 . 0

0 -.2 - . 4 - . 6

X/C Figure 2. Isobars on top surface o f wing (tip region) X = O , a = 120 1 94 I ?

70-

1 ' I

68-

66 -

I

64 -

62-

60-

48 -

46- I 44- 42- 4 2 - %5

38-

Figure 3. Contour pressure plot of the ogee-tip section at a = 8 ' and A = O o Figure 4. Contou pressure plot of the ogee-tip section at Q = 12'and A=Oo Figure 5 . Lift-to-Drag Ratios VS. Angle of Attack W W

0"

io / y p

T

a0

Figure 8. Speed-Power Std. Tips vs. Ogee Tips 16,000 F t .

A

'i

LZ0

T

- 1 Figure 9 . Ogee Tip Configuration R/C vs. KCAS Density Alt . 5549 Ft .

A w Figure 10. Ogee Tip Configuration R/C vs. KCAS Density A h . 9623 Ft .

5,5 W i n g ~ ~ i p Vanes as Vortex Attenuation and

-

Induced Drag Reduction Devices

--

e 1 - 1 . Went+, Jr, and M. G. Nagati Wichita State University Summary Analytical studies have been conducted to exam utilizing wing tip turbines to remove swirl from the wing trailing vortex, and hence reduce the potential for upset of following aircraft. Energy recovery from h e turbines i s also analyzed. A computer routine has been developed to permit rapid parametric studies of various tip turbine designs.

The studies show that the optimum turbine i s a non-rotating set of vanes which reduce swirl and recover energy i n the form of reduced overall configuration induced drag. A specific case study indicates a 23% reduction i n induced drag for a rectangular wing of aspect ratio 5.33, operated a t a lift coefficient at 1 .O.

Introduction The problems associated with the operation o f small aircraft i n the vortex wake of a large aircraft are well documented (Figures 1 and 2 and reference 1).

Many solutions to reducing vortex induced angles have been proposed and bested Most of these however, are achieved at the expense of added drag and hence increased fuel consumption and noise. The present research was undertaken to study the feasibility of a rather novel scheme for diffusing wing tip vorticity, and at the same time recovering energy from the vortex wake.

At least two similar techniques have been tested. The flow straighteners designed by Uzel and Marchman (ref. 2) and the "winglets" developed by Whitcomb (ref. 3).

The present study i s concerned with evaluation of a wind turbine mounted in the center of the wing tip vortex core (Figure 3). The turbine i s designed to re- move the swirl component of velocity from the tip vortex, and to provide rotating shaft torque for conversion to propulsive or stored energy.

Analytical Method The vortex core i s modeled as shown i n Figure 4 based upon Reference 4.

The turbine i s analyzed using the blade element theory of Reference 5. Basic blade e ~ e ~ e n t velocities an angles are s h igure 5, Since the function of the turbine is to reduce upset severity the i s designed t o remove the total swirl velocity at each ra his constrains the local induced angle to a value of one-half the local swirl angle, since half the downwash takes place downstream from the blade element.

It should be noted from the velocity vector diagram that not only i s it possible to obtain a torque-producing force, but it i s also poss i n a direct thrust force component which would appear as a reduced induced drag. The effect has been demonstrated by Whi tcomb's winglets, but evidently the effectiveness of winglets in reducing vortex upset has not been evaluated. Uzel and Marchman evaluated fixed wingtip flow straighteners to reduce the vortex upset hazard, but their design evidently utilizes sharpedged uncambered sections which cannot recover the leading section force which would produce thrust.

A computer program was developed to evaluate proposed designs in the present study. A simplified flow chart of the computational algorithm i s shown i n Figure 6. For simplicity, turbines utilizing constant chord blades were analyzed.

The program was designed to adjust bfade chord until the maximum angle of attack encountered along the span i s between 14.5O and 15O. The effect of the constraint i s to have a design near maximum unstalled lift coefficient condition, i n order to minimize wetted area drag. Computer studies were made at a cost of less than $ 1 per configurution. Design conditions are given in Table 1.

Table 1 . Design Conditions

L i f t Coefficient 1 .o

Aspect Ratio 5.33 Planform Rectangular Core radius .01 (Span) Maximum swirl velocity .8 (Flight speed) Vane skction drag coefficient 0.010 Resu I t s Results of the parametric studies of shaft power output as a function of rpm 7. These data show that shaft power increases and diameter are shown i n Figure with diameter, and that rpm for maximum shaft power decreases as diameter i s increased. Theoretical upper pawer limit occurs when shaft power equals wing Figure 8 presents net power, which i s shaft power minus o r plus the induced power.

drag or thrust power, including blade section drag effects, These data represent a realistic accounting since bl de section drag effects are included. For the va chosen for this study, a maximum net return of 35% of induced power i s achieved for turbine diameters of 32 and 64 vortex core diameters. Thus turbine diameter ratios greater than 32 provide no added benefit. The most intriguing result, however, i s that the optimum rpm i s zero!

I_ From a practical point of view, these results show that for virtually any turbine diameter, the net power recovery i s nearly optimum at zero rpm. Since the i s zero under such conditions, it follows that the significant task of the shaft power tipturbine blades i s t o recover thrust directly. While designs involving blade diameters ratios of 32 do not seem practical, designs with diameter ratios i n the range of 2 to 8 may be quite feasible.

An additional computer run was made with zero rpm, turbine diameter ratio equal t o 8, and vane section drag coefficient increased from 0.010 to 0.013.

A l l other parameters were retained as given in Table 1 , This run showed that an induced power recovery at 23% was possible. This i s believed to be a very realistic set of design conditions.

Blade twist distributions for selected configurations are shown in Figure 9.

These results indicate that twist requirements pose no extreme fabrication problems.

Concluding Remarks The present analysis has many limitations. Some of these are debcribed and discussed below: 1. Vortex core rollup i s not complete one chord behind the trailing edge.

Therefore the assumed vortex velocity model i s only approximate.

2 The present analysis does not account for mutual aerodynamic interference between the vanes and the wing. The vane l i f t will certainly produce induced velocities which will influence the main wing lift and hence vortex distribution. A more sophisticated analysis would include mathematical modeling of main wing and tipvane vortex.

3. The present analysis does not account for off-optimum performance. It i s reasonable to expect that operation at lower wing lift coefficient will adversely influence performance, since blade twist will no longer be optimum. The design criteria should protect against stalling of the vanes under such conditions, however.

4. The effect o f the tip vanes i s to replace a single concentrated vortex with a series of vortices emanating from the vane tips. N o analysis has been made of the t r a j e c t ~ y of this new vortex system, If the new vortices coalesce, i t i s to following aircraft might not be reduced.

possible that the upset ha Conclusions 1 . Feasibility studies indicate that tip mounted multi-vane turbines can recover energy from a wing vortex wake, while simultaneously reducing the vortex swirl and presumably the upset hazard to following aircraft.

2. The studies show that a non-rotating array of vanes properly twisted w i l l provide maximum net energy recovery, in the form o f vane thrust.

3. Practical designs from the present study should be evaluated by wind , tunnel tests to determine actual performance gains, as well as penalties for off- design operation.

References 1. Roberts, L. : On Vortex Wake Alleviation. In "The Future of Aeronautics, I' Proceedings of NASA University Conference on Aeronautics, University of Kansas, Lawrence, Kansas, October 1974.

2. Bower, R.E.: Opportunities f o r Aerodynamic-Drag Reduction. In "The Future of Aeronautics," Proceedings of NASA University Conference on Aeronautics, University of Kansas, Lawrence, Kansas, October 1974.

JON. and Marchman, J.F. 111: The Effect of Wing-Tip Modifications Uzel, 3.

on Aircraft Wake Turbulence. Report No. VPI E-72-8, Virginia Polytechnic Institute and State University, July 1972.

Scheiman, and Shivers: 'Exploratory Investigation of the Structure of the Tip 4.

Vortex o f a Semispan Wing for Several Wing Tip Modifications, NASA TND-6101.

Dommsch, D.O., Sherby, S.S. and Connolly, T.F., Airplane Aerodyna- 5 .

mics, Pitman Publishing, 1967.

figure 3 . Wing-Tip Devices

L

c 3 .

b

$ 3

s :

I /

b r

ASPECT RATIO, cL, SPAN, AREA, SPEED.

ASSUME BLADE AREA

i I ; 7

CALCULATE SECTION LIFT c WRITE RESULTS BETA =f (RADIUS) , THRUST, TORQUE, POWER Figure 6. Tip Turbine Design Computer Program 21 1 Figure 7. Shaft Power Ratio 21 2 d Figure 8. Net Power Ratio 21 3 0 0 tr Q) c IIC 21 4 University of Kansas Introduction The drag of a wing may be classified as essure drag, i nduced drag and skin friction drag. The induced drag i s Induced by the vortex system set up around the finite three dimensional wing. By decreasing the strength of the trailing vortices, the induced drag may be reduced.

The basic problem i s to decrease the vortex without increasing the pressure and skin friction drag so that the total i s decreased, or maximizing L/D.

The actual induced drag i n pounds is: I nb qe By changing planform e may be varied up to the value for the ideal o f the elliptical span loading. However, a two dimensional equivalent wing requires a maximum value o f circulation of only rr/4 that of the maximum for an elliptical wing loading.

Thus, i f the shed vortex could be reduced by this amount, the induced drag warld be reduced 21 percent. Since induced drag accounts for 25 percent to 40 percent of the total aircraft drag, this would mean a total reduction of 5 percent to 8 percent over the eltiptical loading. A number of methods such as wing,tip end plates, tip tanks and winglets have been used o r tested to provide an effective increase i n aspect ratio and achieve a more two dimensional wing loading. In order to control the wing tip vortices the basic vortex characteristics need to be considered.

Basic Vortex Characteristics The basic characteristics of the vortex are shown in the tornado (1) , Figure 1. The core flow region and the free vortex region are clearly evident.

Laboratory investigations(2) have shown that the strength of an unconfined vortex such as the wing tip vortex is a direct function of the vorticity present and the sink pressore and the area to create a core flow to organize the vorticity into a vortex.

The circulation type vortex i s shown in Figure 2 , large diameter with little axial flow.

Introducing an axial pressure differential, Figures 3, 4, and 5 show the development into a strong compresible flow vortex, Fi gure 6. Figure 7, the circulation i s continued with se in axial pressure differential a Vortex breakdown occurs e Figure hows the continuation of the compressible flow vortex with pressure djfferent~al only, The rotation of the cage has been stopped.

vortex i s unstable and wanders around. Figure 9 shows the pressure trace at a fixed point as the compressible core, Figure 6 , moves over and around the point, Idealizing the compressible flow vortex the core flow, pressure and density are shown in Figure 10. Figure 11 shows the formation of two compressible flow vortices in Figure 12. Using neutrally buoyant helium bubbles the compressible core shown flow is evident in Figures 13 and 14. Continuing circulation and decreasing the axial pressure differential, the core flow breaks down, Figures 15 and 16.

The vorticity shed i n producing lift and the sink provided by the negative pressure region on the upper surface must both be minimized and/or neutralized to decrease the induced drag. Laboratory tests have shown that the introduction o f pressure i n the core will stop the care flow and dissipate the vortex. An obstruction screen or splines can be introduced into the core flow to attenuate the vortex.

Counter vorticity likewise i s effective i n reducing the vortex. In a number of tests the introduction of turbulence by various means has set up instabilities in a i t s decay. Thus, there are a number of avenues available vortex which has hastened for some measure of vortex control.

Wing

-

The basic object is to maximize the wing L/D for a particular operating condition for a given aircraft. Most o f the research to date on the wingctip vortex has been conducted for the purpose of attenuating the vortex downstream of the aircraft, Figure 77. A drogue device properly positioned downstream of the wing tip causes breakdown, Figure 18. A jet engine simulator at the tip with a high-energy jet blast produces the same results.

To increase L/D by decreasing the induced drag the two basic parameters of the vortex must be controlled: 1. The vorticity shed by the wing must be minimized.

2. The low pressure region on the top of the wing must be blocked from the shed vorticity. It i s the wing-tip vortex core flaw (or deficit flow) that i s largely responsible for producing the induced drag.

The vorticity may be reduced by plates, tiptanks, winglets and counter vortex flow which effectively increase the aspect ratio of the wing. The general effect of these on L/D are shown in Figures 19, 20, and 21. It w i l l be noted that 21 6 ck a t which the device offers the I rgest advan~~ge art~cuiar angle of expected that any one of these could be optimized further as to size, of the wing, shape, angle and contour for a given angle of Efforts to use a j e t have been directed at reducing vorticity downstream.

Figure 22 shows the effectiveness of various jet strengths on the dissapation of vorticity. Obviously the upstream jet i s not effective in reduction of o f the downstream jet may be pbrt of the answer. Again th proper use location and strength must be optimized for maximum L/D.

The Method of Approach The method o f approach for optimization of L/D through minimizing induced drag should be through a detailed flow study together with force, pressure and vorticity measurements. Flow visualization with neutral helium bubbles, Figure 23 and 24, provides an excellent means of observing the effects of configura- tion changes. A systematic wind tunnel investigation of a large number of configura- tion changes should be made. The study should explore a l l avenues which appear promising us the study progresses even though i t may lead to the rebirth of the bi- plane o r triplane.

References Muirhead, V.U., "Compressible Vortex Flow," AIAA Paper No. 73-106, 1 .

1 l t h Aerospace Science Meeting, New York, New York, January 1973.

2. Eagleman, J.R.; Muirhead, V.U.,; Willems, N ., Thunderstorms, Tornadoes, and Building Damage, Lexington Books, D.C. Heat% and Company, Lexington, Mass. , 1975.

Figure 1 . Tornado Figure 2 . Circulation Vortex 21 8 Figure 3 e Introduction of Axial Pressure Differential Qo Figure 4 I Ear I y Vortex Deve lop 21 9

Vortex Development - Spiral Vortices

Figure 5.

Figure 6 , Compressible Vortex Flow

Figure 7 e Vortex Breakdown - Circulation with Decreasing

Axial Pressure Differential Figure 8, Vortex Sustained by Axial Pressure Differential 22 1 1.0 0.9 0.8 0.7 E .

PA 0.6 0 . 5 0 5 10 15 20 25 30 35 40 45 Mil I iseconds Figure 9 . Pressure Traces 1.0 .5 M 1.0 P PA .5 1.0 P

-

0 .

1.0 2.0 3.0 r Figure 10. Compressible Vortex Core Conditions Figure 1 1 . Two Vortices Forming Figure 12 I Two Strong Compressible Vortices a , L a Is)

.-

U

-

m m a *- e a , E Y m X a , c b

>

u-l 1 " - E

.-

1 1 Figure 17. Vortex System Figure 18. Attenuation Methods Figure 19. L/D of Wing Tip Configurations Figure 20, L/D of Wing Tip Configurations Figure 21. L/D of Wing Tip Configurations

CJ - JET MOMENTUM COEFFICIENT

Figure 22. Jet Effect on Wing Tip Vortex

Figure 23. Rounded Wing Tip - Flow Visualization

Figure 24, Wing Tip Plate - Flow Visualization

4, INT G 6 . ternal Nacel I e ag and Interference R. D. Neal, Gates Learjet Corporation 6.2 Instal lation Drag ConsQderations 0 ther than External Nacelle and Interference Drag as Related to Turboprop and Turbofan Engines G . Burnett, Garrett AiResearch Manufacturing Company o f Arizona 6.3 Nacelle Drag Reduction: An Analytically Guided Gcperimental Program F. Smetana, North Carolina State University 6.4 Cooling Drag Associated with Genera1 Aviation Propulsive S ys terns

E . J . Cross, J r . , Mississippi State university

6.5 Propellers of Minimum Induced Loss, and Water Tunnel Tests of Such a Propeller E. E. Larrabee, Massachusetts Institute of Technology

Preceding Page Blank

6 , l ~ v e r v i e w of External Nacelle and Interference Drag Ronald D. Neal Gates Learje t Corporation Introduction The purpose of this paper i s to provide a written outline record of an oral presentation given at the General Aviation Drag Reduction Workshop.

Historical Review of Mu1 ti-Jet Engine Installations Airplane performance i s achieved thru a Combination of aerodynamics and propulsion. In terms of the propulsion system, airplane configurations-be they powered by propellers or jets--are developed around the characteristics of a specific engine. For t h i s reason, the integration of the powerplant and the airframe is truly the cornerstone of the aircraft design. The introduction and continued development of the turbine engine has only served to emphasize the importance of achieving a successful engine/airframe interface.

The beginning of the jet age took place on August 27, 1939, when the German Heinkel HE-178 research airplane made i t s f i r s t flight. This airplane was powered by a single gas turbine engine having a thrust of about 1,100 pounds.

The next jet airplane--and the first twin engine jet-was another Heinkel design, the He-280. Powered by two 1,320 pound thrust engines, this airplane made i t s initial flight in April 1941.

The next jet fo fly was the Messerschmiti Me262, which was powered by two 1,850 pound thrust axial flow turbine engines, with the f i r s t flight occurring in July 1942. The Me262 certainly ranks as one of the most advanced aircraft designs to be developed during the Second World War and it also has the distinction of being the first jet aircraft to reach operational status.

By h e end of the war, the German aviation industry had developed several jet aircraft designs. Examples of actual production, flight hardware include the single-engine He-162 fighter and the twin-engine AR234 bomber.

The first allied jet to fly was the British Gloster E28/29. This airplane, powered by a single 860 pound thrust gas turbine engine, designed by Frank Whittle, made i t s initial flight i n May 1941.

Preceding Page Blank

The United States entry into the j e t era took place on October 1, 1942, when -59A "Airacomet" took to the air.

the iwin-engine he second American jet to fly was the single-engined Lockheed XP-80, with this flight taking place i n June 1944. Neither the P-59 or the P-80 were to see combat in World War I!, however, i n November 1950, in the skies over Korea, an F-80 became the winner of the f i r s t all-jet aerial combat by downing a Russian built MIG-15.

The post-war years ushered in a whole new era in aircraft design. Examples of some of the multi-engine airplanes flown i n this period include the twin-engine 8-43, the three-engined B-51, the four-engined 8-45 and 8-46, the six-engined B-47 and 8-48, and the eight-engined B-52.

I n July 1949, the four-engined deHavilland Comet 1 made i t s f i r s t flight and the dawn of commercial jet transportation had begun, America's first jet trans- port was the Boeing 707 which made its maiden flight in July 1954, with the f i r s t 707 transatlantic service beginning i n October 1958. The introduction of j e t service on this transatlantic route reduced the flight time from twelve hours to seven hours, I n May 1955, the French entered the commercial transport field with the Sud-Aviation Caravelle. The Caravelle, with i t s two jet engines mounted on the aft fuselage, represented a design innovation that i s s t i l l in vogue some twenty years later. The commercial aft-engine j e t transports that have been developed thru the years include the following: Sud-Avia tion Caravel f e BAC-I 1 Douglas DC-9 Fokker F-28 TU-134 Yak-40(3-engine) Yak-42 (3-engine) Boeing 727 Hawker Siddeley Trident (3-engine) TU-154 (3-engine) I I yushi n 1 1-62 (4-engi ne) BAC VC-IO (4-engine) It i s perhaps of historical interest to note that the original patent for the in November 1951, $and w s entitled "Improvemenk in Caravelle design was filed

Aeroplanes Propelled by Several Jet Engines ."

Development of Business Jet Aircraft With the successful introduction and acceptance of commercial j e t transports i t was only a question of time until the performance potential of turbine power was applied to the general aviation airplane.

he origin of the business jet can be traced to the four-place, Fr orane-Saul nier S760, which first flew i n mid-1954. However, a 955 attempt by ircraft to market this airplane i n North America con best be described as unsuccessful.

The next airplane to enter the small jet transport arena was the Lockheed Jetstar, with the original twin-engined prototype flying i n September 1 9 s . The Jetstar was originally designed for the military market i n response to the for a small jet transport with the eventual outcome of this effort being the four-engined C-140.

The next airplane to come along was the North American Sabreliner flying i n September 1958, as an entry into the military UTX competition for a trainer category airplane.

The third small transport took to the air in February 7959, when the four- engined McDonnelI Model 220 flew, with this airplane also competing for the UCX contract .

In the final analysis, the Jetstar won the UCX race, the Sabreliner took the UTX contract and McDonneil dropped the Model 220 program.

The next period of activity i n this field took place i n 1962, when the first deHaviIIand DH 125 flew. The following year, 1963, produced a bumper crop of airplanes with the first flights of the Jet Commander, the French designed Mystere 20, and the Lear Jet taking place. The swept-forward wing German Hansa Jet and the Italian PD808 made their first flights i n 1964. A new era in big business jets began when the Grumman Gulfstream I1 made its initial flight i n 1966. The latest business jets to join the field include the Cessna Citation, the Falcon 10, and the Corvette.

All of these business jets (with the exception of the MS760 and the McDonnell 220) are of the aft fuselage mounted engine configuration.

While the large commercial transports and the smaller business jets are similar in configuration, there is a difference between the two designs. Specifically the aft-engined transport aircraft tend to have the nacelles located well aft of the wing trailing edge, for example the DC-9 and 727. In the case of the smaller airplanes, the nacelles are located quite close to the wing and in several designs the nacelles overlap the wing. Because of the proximity of the nacelles to the wing, the business jet offers some challenging design problems i n terms of achieving a minimum drag configuration.

Also, the trend in business aircraft design hos been towards the incorporation of high bypass fan engines. These engines, with their larger physical size, make it ounted engine arra that i s compatible with siness jet design, i n the years ahead.

A good example of the size impact of a turbofan engine i s shown by the Learjet testbed airplane which incorporated the General Electric CJ610 turbojet engine on one side and the AiResearch TFE 731 turbofan engine on the other side.

Current Business Jet Engine Installations Slides number 44 through 50 provide installation photos of various turbojet/ turbofan aft fuselage mounted engine arrangements on current business jet designs.

Comments an some of the aerodynamic aspects of the installations were offered.

Aft-Engine Nacel le Drag Considerations Viewgraph 1* presents sketches of a long fan duct and a short fan duct nacelle considered for the FTE 731 installation on the Learjet Model 35/36.

Viewgraph 2 presents a typical nacelle configuration trade-off h a t can be made for various design studies.

Viewgraph 3 presents wing pressure distribution as affected by nacelle position (data from Reference 9).

Viewgraph 4 presents typical nacelle drag characteristics for a turbojet engine installation and Viewgraph 5 shows the drag characteristics for a turbofan nacelle.

Viewgraphs 6 and 7 show some nacelle geometry configurations.

Viewgraphs 8 thru 12 present some nacelle drag results obtained with various nacelle locations.

Third- Engine Location For a three-engined airplane there are two rather obvious locations for two of the engines, either on the aft fuselage or wing mounted. As for the third engine, there are also at least iwo options with examples being the S-duct (727, L-IO? 1) or the straight-through duct (DC-1 0 ) .

There appears to be little published information on comparisons between these two types of installations. b e i n g has reported (Reference 18) that their studies have shown that the weight/performance trade between the S-duct and straight-duct i s about even. Design studies conducted by Lockheed have shown that the S-duct offers better "Viewgraphs not included i n written version for proprietary reasons.

ance than a s~ajght- duct^ O n the other hand, onnel I-Douglas studies have identified the performance improvements of the straight-duct over the S-duc t a he final choice for the type of third engine installation i s not completely clear, however, in terms of numerical numbers the S-duct i s the winner. I f the weight and drag are i n fact an even trade between the two concepts, then other installation factors w i l l dictate the final selection. I n fact, as part of the Advanced Transport Technology (ATT) studies, United Airlines (Reference 20) preferred the S-duct from an engine maintenance viewpoint due to the lower engine position.

I n the small transport c.ategory of aircraft, both the proposed Cessna 700 and the Falcon 50 have elected to utilize the S-duct arrangement.

In terms of an historical viewpoint the Martin XB-5T, which flew i n October 1951, had its third engine located i n the rear fuselage with inlet air being provided via an S-duct configuration.

From the civil aviation standpoint, Sud-Aviation applied in December 1951, for a patent covering "Improvements i n Aircraft Equipped with a Propelling Motor at the Rear" with this patent covering various S-duct configurations.

Slides number 62-67 provide illustrations of various third-engine installations.

Viewgraphs 13 thru 15 provide additional information of S-duct configurations.

R & D Study Recommendations There appears to be a need and requirement to investigate the inter- a) ference drag of nacelle configurations mounted on the aft fuselage with specific emphasis on configurations having the nacelle i n close proximity to the wing.

For high bypass ratio turbofan engines the drag interference problem b) of the short cowl nacelle on the aft fuselage should be examined.

There appears to be a need for published research information on the c) aerodynamics of S-ducts and other third-engine instal lation aerodynamics.

Reference Material The foll owing reference material presents some sources containing information relating to the overall subject of aft-engine installations and basic nacelle design considerations, 2 39

rd J . , " Deep-S mic Character-

ence on Aircraf roblem, May 1965, NASA SP-83, pages 1 13-1 21 .

10-12, 2. Ray, Edward J., and Taylor, Robert T., "Effect of Configuration Variables on the Subsonic Longitudinal Stability Characteristics of a High-Tail Transport Configuration .I' NASA TM-X-1165, October 1965.

3. Taylor, Robert T., and Ray, Edward J . , "Factors Affecting the Stability of

T-Tail Tranports .I' Journal of Aircraft, July-August 1966, pages 359-364.

4. Taylor, Robert T., and Ray, Edward J., "A Systematic Study of the Factors Contributing to Post-Stall Longitudinal Stability of T-Tail Transport Con- figurations ." NASA TM-X-56882, November 1965.

5. Thomas, H.H.B.M., "A Study of the Longitudinal Behavior of an Aircraft a t Near-Stall and Post-Stall Conditions .I' AGARD Conference Proceedings No.

17, September 20-23, 1966, pages 729-769.

6 . Kettle, D . J., and Kerby, D. A., "Low-Speed Wing Tunnel Tests on the Effecis of Tailplane and Nacelle Position on the Superstall Characteristics of Transport Aircraft." R.A. E, R&M 3571, August 1967.

7 . Shevel I , Richard S., and Schanfele, Roger D., "Aerodynamic Design Features of the DC-9." AIAA Paper 65-738, November 1965.

8. Waaland, 1. T., and Curtis, E . J . , "Gulfstream I1 Aerodynamic Design."

SAE Paper 670242, April 1967.

9. Lobert, G . , and Thomas J . , 'I Engine Airframe Ingegration Problems Peculiur

to Aircraft Configurations with Nacelles Mounted Above the Wing .'I Paper presented at 31st meeting'of *he AGARD Flight Mechanics Panel, Gottingen, Germany, September 13-1 5, 1967, 10. Putman, Lawrence E . , and Trescot, Charles D . , Jr., "Effects of Aft-Fuselage- Mounted Nacelles on the Subsonic Longitudinal Aerodyanmic Characteristics of a Twin-Turbojet Airplane. 'I NASA TND-3781, December 1966.

Aoyagi, Kiyoshi, and Tolhurst, William H., Jr., "Large-Scale Wing-Tunnel 11.

Tests of a Subsonic Transport with Aft Engine Nacelles and High Tail .'I NASA TND-3797, January 1967.

Neal, Ronald D., "Correlation of Small-Scale and Fuli-Scale Wing Tunnel 12.

Data with Flight Test Data on the Lear Jet Model 23." SAE Paper 700237.

13. William, P. R. G., and Steward, D. J., "The Complex Aerodynamic Inter-

ference Pattern Due to Rear Fuselage Mounted Power Plants." AGARD Conference Proceedings No. 71, Aerodynamic Interference, September 1970.

14. Callaghan, J. T., Doneison, J. E . , and Morelli, J. P., "The Effects on

Cruise Drag of Installing Re-Fan - Engine Nacelles on the McDonnelI Douglas

DC-9." NASA CR-121219, May 1973.

Aft-Fuselage- 5 , rasnport Configu 78, March 1975,

. , and Sigalla, Armand, "The roblem of Installing a

6 .

High Bypass Engine on a Twin Jet Transport Aircraft." AGARD Conference Proceedings No. 124 on Aerodynamic Drag, April 1973.

17. Williams, P. R. G., and Steward, D. J., "An Aircraft Designer's Review of Some Airframe and Engine Ingegration Concepts .It Paper presented a t 1 s t International Symposium on Air Breathing Engines, June 19-23, 1972.

18. Goodmanson, Lloyd T., and Schul ta, William H., "Installation and Inte- gration of Transonic Transport Propuision Systems ." SAE Paper 71 0762.

19. Hawkes, J. E., "Development Status of the L-1011 TriStar." SAE Paper 71 0755.

20. "Assessment of the Application of Advanced Technologies to Subsonic CTOL Transport Aircraft. I' NASA CR-112242, April 1973.

Nacelle Design 21 * Saylor, James M. and Smith, Robert E . , "Internal and External Aerodynamics of the C-141 Nacelle. I' A I M Paper 65-604.

22. Allison, H. B., and Leslie, H. R . , "Installation Considerations for High Bypass Ratio Turbofan Engines .'I AIAA Paper 67-390.

Viall, W. S . , "Aerodynamic Considerations for Enginer Inlet Design for

23.

Subsonic High-Bypass Fan Engines. 'I SAE Paper 660733.

24. Frazier, G. T., "Aerodynamic Considerations for Engine Exhaust Design far Subsonic High-Bypass Fan Erg ines." SAE Paper 660734.

25. Viol 1, W. SI, "The Enginer Inlet an the 747. 'I ASME Paper 69-GT-41.

26. Saylor, J. M., and Hancock, J. P., "C-5 Engine Inlet Development."

ASME Paper 69-GT-52.

27. Hancock, J. P., and Hinson, B. L., "Inlet Development for the L-500."

A I M Paper 69-448.

28. Rohling, Walter J., "The Influence of Nacelle Afterbody Shape on Airplane Drug. "

29. Leynaert, Jr , , I' Engine Installation Aerodynamics, I' paper given a t AGARD

Lecture, Series No. 67, Prediction Methods for Aircraft Aerodynamic Char- acteristics. AGARB-LS-67, May 1974.

30 Tunnel Investigation of Jet-Wake Effect celle rnterf@r@nce Drag of a Subsonic Transport. 'I NASA , August 1968.

Patterson, James C., Jr., and Flechner, Stuart G., "Jet-Wake Effect of a 31.

High-Bypass Engine on Wing-NaceI l e Interference Drag of CI Subsonic Trans- port Airplane ." NASA TND-6067, November 1970.

Raney, D. J., Kurn, A. G., and Bagley, J. A,, "Wind Tunnel Investigation 32.

of Jet Interference for Underwing Installation of High Byposs Ratio Engines. I' ARC C.P. No. 1044, March 1968.

Kurn, A. G . , "A 'Further Wind Tunnel Investigation of Underwing Jet Inter- 33.

ference." ARC C. P. No. 1156, April 1969.

Powered Nacel le Testing 34. Fasano, A., Erlandsen P., and Barrett, D . , "The Powered Nacelle as an Experimental Tool .'I AIAA Paper 70-636, June 1970.

Welge, H. R., and Ongaroto, J. R . , "Powered Engine Simulator Porcedures 35.

and Experience for the DC-10 Wing Engine at High Subsonic Speeds." AIAA Paper 70-590, May 1970.

Grunnet, James L . , . "Designing Jet Aircraft Wing-Tunnel Test Programs with 3 6 .

Propukion System Simulation." J. Aircraft, Vol. 8, No. 6, June 1971 List of Slides 1 . He-178 Single-Engine Jet Airplane 2. He-280 Twin-Engine Jet Airplane 3. Me161 Twin-Engine Jet Airplane 4. He-162 Single-Engine Jet Airplane AR 234 Twin-Engine Jet Bomber 5.

6. Gloster E28/39 7. Bell XP-59 8. Lockheed XP-80 9. buglas B-43 10. North American 8-45 1 1 . Convair B-46 12. Boeing B-47

. Martin 8-48

14. Martin B-51 15. Boeing B-52 16. Boeing 707 17. Caravelle 18. DC-9 19. Fokker F-28 20. BAC-111 21. Yak40 22 * 23.

24.

25.

6. Jetstar 27. Sabre1 i ner 28. McDonnell 220 29. DH-125 30. Jet Commander 31. Falcon 20 32. Learje t 33 t Hansa Jet 3 4 . PD808 Gulfstream I1 35.

Citation 36.

S N6OO 37.

Fa1 con 38.

39. Dc-9 40.

41 . Learjet 35/36

Lear'et Test Bed Airplane 42.

43. BIan f, Slide 44. Front view of JTl5D Citation Front view of TFE 731 on Falcon 10 45.

Front view of TFE 731 on Learjet 46.

Front view of CJ610 on Jet Commander 47.

Aft view of TFE on Falcon 10 48.

Aft view of TFE 731 on Learjet 49.

Aft view of Larzac on Falcon 10 50.

Side view of JT 150 on Citation 51.

Side view of JT15D on SN600 52.

53. Side view of Larzac on Falcon 10 Side view of 05610 on Learjet 54.

Side view of JT12 on Sabreliner 55.

Side view of CF700 on Sabreliner 56.

Side view of CF700 on Falcon 20 57.

58. Aft view of CF700 on Falcon 20 Side view of TFE 737 on Learjet 59 * Side view o f TFE 731 on Falcon 10 60.

Black Slide 61.

Boeing 727 62.

63. Lockheed L -1 01 1 64. Yak 40 65. Falcon 50 66. Dc-IO 67. Martin B-51 Blank Stide 68.

i s t of Viewgra~hs 1 . Nacelle Configurations c t TradeOoff 2, 3. Effect of Nacelle Position on Wing Pressure Distribution 4. Turbojet Nacelle Drag Characteristics 5. Turbofan Nacelle Drag Characteristics 6 . Nacelle Location 7. Nacel I e Location 8. Effect of Nacelle Incidence 9 . Effect of Nacelle Rotation Effect of Nacelle Position 10.

1 1 . Effect of Nacelle Rotation Effect ofNacel le Incidence 12.

Design Comparison of the L-1011 and DC-10 Third-Engine Installations 13.

14. Comparison of Boeing 727 and Lockheed L~1011 Duct Contours 15. Sud-Aviation 1951 S-Duct Patent

G . A. Burnett

Garrett AiResearch Manufacturing Company of Arizona Introduction Considerable effort i s presently being expended by NASA, various univer- sities, and industry to improve and develop technology in many areas directfy appli- cable to general aviation aircraft design. One of these major areas i s directed toward new airfoil designs for improved lift-to-drag-ratio characteristics for improved climb and cruise performance. Another i s directed toward high-l ift-device improvements that could open the door for increased wing loading design criteria, thus reducing wing area and cruise drag. The results of these programs will undoubtedly provide some significant aerodynamic improvements when the research and development work has been completed; however, the testing, proving, and optimization of most of these concepts are still i n the early-to-moderate stage with respect to being introduced into production general aviation aircraft.

With this i n mind, i t would appear advantageous to approach the problem of improved aircraft performance and/or drag reduction along a t least two parallel paths which consist of new technology development and identification of areas where potential improvement with existing technology could be attained. The latter would also tend to complement advanced technology.

One such area i s the drag penalties associated with propulsion system instal- lation. Typically, at representative cruise operating conditions, the total installed drag of a turbofan engine installation can effectively amount to between 10 and 15 percent of the total aircraft drag. Similarly, a furboprop engine installation can amount to between 20 and 40 percent of the total aircraft drag. As a starting point, sttaight jet and turboprop engine instal lation some of the specific areas associated with and, thus, improved aircraft system per- have been outlined where drag reductions formance can be obtained.

Discussion Before the subject of drag reduction can be addressed, an accounting pro- cedure for evaluating the propulsive effort must be defined. For the straight jet engine installation, this is a re.latively simple procedure, as shown i n Figure 1 .

t L

ro

w c Q, L 0)

.-

LL It II It II II 11 P

1 "

LL.

n u .

a

u . P

n

w With the use of accounting procedures that have been accepted where drag is defined a s the s u m ~ a t ~ o n of forces acting on the outside of the stream tube bounding the flow that passes through the complete engine and thrust i s defined as the summation of forces on the inside of the stream tube, the complexity of thrust and drag accounting becomes relatively simple.

Obviously, this same exact procedure cannot be applied to a propeller powered installation, since the stream tube or slip stream now has moved from the inside of the engine to the outside. However for a turboprop engine insta!lation, an extension of the basic straight j e t accounting procedure may be established as shown i n Figure 2.

The purpose of defining an accounting procedure i s twofold. First, i t provides the means of completing a preliminary performance assessment of one engine installation with respect to another, which i s an obvious requirement for aircraft performance analysis and trade-off studies; and secondly, i t provides a method to identify areas This procedure has apparently not been as fully utilized of potential improvement.

on propeller installations as straight jet installations. This i s indicated by the lack of design guidelines and installation aerodynamic trade-off data. This may be attributed in part to the fact that propel ler-powered aircraft engine installations come i n many variations, whereas straight j e t engine installations are fairly standard i n terms of comparing one installation to another, independent of thrust or application.

Air-Intake Design Considerations All turboprop and straight j e t aircraft propulsion system installations have from the free stream into the engine. Most primary air intakes for directing airflow installations utilize secondary air intakes for providing cooling and ventilation air- flows to various components and hot sections of the engine. The design considerations i n terms of sizing, design-point selection, location, and shape can significantly affect the propulsive effort of the propulsion installation (net thrust, nacelle drag, and additive drag).

The design objective for most business j e t intake systems i s minimum length for weight and surface area considerations while maintaining a high drag-rise Mach number, low spillage drag characteristics, and high ,total pressure recovery with low flow distortion to the engine With the advent of modern high-bypass-ratio turbofan engines (high flow per unit frontal area and increasing maximum diameters), this objective has become quite a challenge to the aerodynamicist. I f the intake sizing i s too large for the required engine airflow (low mass-flow ratio), flow spillage L Y

z

E uj n I- I o W -I I& U UJ K 0 Q E I w U n t- a

z

J I v) ct u .

I n w LL c- d

-

w w [L [r: I L.

cc

Y

c W L.

W m .A t E -i w u .

u w

a

z

Z w

z

(Y 0)

.-

u .

(3 h m a

a

D .A li II If II I J ii 1 1 II II II II v) --.I a F II ..

Lu c I 3 211 results which can lead to flow separation. I f the forebody sha e {fineness ratio) i s not adequate, supersonic expansion can occur which m a y result in flow separation. I f the inlet lip (from the h ~ g h t ~ g h t to h e throat) and internal diffuser characteristics are not considered, excessive additive drag can result.

Up to now the NACA Series I profile has been used for most forebody air- intake designs; but at low mass flow ratios, excessive spillage drag can result due to the high local flow angle a t the inlet l i p or highlight. This i s especially true of modern high-bypass-ratio turbofans used on general aviation aircraft where fixed- geometry- air intakes are used predominantly. The air-intake throat i s sized for good cruise diffuser performance, but the static takeoff conditions require generous highlight- to-throat-area-contraction ratios to preclude flow separation during static ground and crosswind operation. A s a result, during some operating conditions (speed and engine power setting), extremely low mass flow ratios can result. While operating i n these conditions the stagnation streamline can be located we1 I wiihin the air intake to the inside of the highlight, which will require the flow on the outside of the streamtube (spillage flow) to rapidly accelerate and expand around the highlight within the for- ward region of the cowl. I f the flow separates, the effect of the suction pressure loss

reduces the lip suction force and , thus, increases the additive drag in addition to

the basic pressure drag of the nacelle. Some recent studies have suggested that the problems associated with low-mass-flow air-intake operation may be alleviated by incorporating forebody profile shapes similar to those being investigated for super- the modified forebody critical airfoils-the principle being that the suction pressure on i s retained well beyond the point where suction pressure collapse occurs on o shapes Series 1 profile.

As shown in Figure 3, the reduction in additive drag from a NACA Series 1 forebody and a modified supercritical forebody i s indicated as: Mass Flow Ratio CD Spillage, Based on Frontal Area 0.6 -43% 0.4 -77% With turboprop engine instal lotions, the problems associated with air-intake design can become more of a challenge than that of straight jets. This can be attri- buted to propeller slipstream interaction effects, which complicate accurate local flow field &finition. As a consequence, the air intakes on most propeller-powered aircraft are oversized to offset the uncertainties, thus resulting in high additive drags, increased surface areas, and propeller blockages. In addition to the basic E

.-

t

E

u l C v c rn

.-

v)

d

(u Y d

e

z

c I Q L

Y

a

a m

a

(3 U u) E iz n u .

+

cc c + n g.

!I

a

c r

u1 v)

a

m

*

0- u .

Y a \ t x w U I U w x a

t i

=

p c u a a a r 25 1 drags associated with the air intakes, the arasitic drag resulting from local flow separations on the nacelle due to prediffu Turboprop Exhaust- Duct Arrangement Some turboprop engine installations offer options i n the approach to designing the required exhaust duct and cooling systems. When these optio in terms of aircraft constraints, cost, weight, and performance sh to assess the best configuration for the engine installation and, thus, the total aircraft system .

Figure 4 shows three possible exhaust-duct configurations that may be con- sidered for a typical turboprop aircraft installation. As shown, the three configurations consist of a straight duct that has been designed to minimize internal pressure losses (no bends;; minimum length), to provide maximum use of the j e t thrust, and to minimize frontal area or blockage.

The second duct i s a typical compromise that could be encountered on some installations. Like the straight exhaust, i t has been designed to utilize the available j e t thrust, but at the expense of additional internal pressure l o s s and external drag.

The third duct illustrates a configuration where the designer may consider minimizing external drag and frontal blockage at the expense of utilizing the engine exhaust jet energy.

To provide insight as to impact on propulsive effort of the three exhaust-duct configurations considered a simple performance assessment i s shown that considers the relative effect of each configuration with respect to the power attainable with an uninstalled specificatisn engine. The result obtained from this parametric analysis i s unique for each exhaust-duct area considered ,with respect to internal pressure loss and external drag.

As expected, the straight duct configuration results i n the smallest power loss (approximately 1.5 percent). The difference between the compound side exhaust (optimum area) and the straight duct (optimum area) i s approximately 5.0 percent, which i s attributable directly to external drag and internal pressure-loss effects on to be the engine. The optimum area stub side exhaust performance was estimated approximately 8 percent lower than the straight exhaust duct.

In terms o f airplane drag, the difference between the optimum straight duct stub side exhaust design represent 30 to 35 Ibs drag differential at a design and the typical cruise operating condition.

previously indicated for exhaust-duct trade-offs, turboprop engine cooling requirements (compartment vent~lation and oil cooling) provide some design alternatives.

Most systems use either full ram systems, which are dependent upon recovering kinetic energy from the propeller slipstream or free-stream velocity, or augmented systems using the kinetic energy of exhaust velocity to provide an eductor. Both systems have advan- tages and disadvantages.

A t static or low-speed operating conditions, where the free-stream kinetic energy i s low, eductor systems can provide the augmentation necessary to obtain the required cooling flows; however, the optimization o f an eductor system requires a com- plete parametric analysis a t the design point and off-design operating conditions to fully assess the interaction of the interrelated flows and the effect on propulsive effort.

I n comparison, full ram systems are simpler to analyze due to the elimination of the interacting flow fields. Improperly sized eductor systems can result in significant engine power loss and ram drag at normal cruise operating conditions.

As indicated previously, full ram systems are less risk to design than flow- augmentation systems. Proper designs can be obtained that result i n minimum per- formance loss to the aircraft i f proper design criteria are followed for air-intake sizing, internal diffuser design, and flow control employed for cruise operation where the cooling flow requirements are low.

Figure 5 shows the cruise power loss as a function of flow control area ratio for a full eductor cooling system and an isolated ram cooling system design. The points at 100 percent area ratio show the power loss i f no flow control i s used. As indi- cated, the power l o s s of the full ram system amounts to approximately 6 percent ( o i l cooler plus compartment ventilation), whereas the eductor system cruise power loss i s only 2 to 2.5 percent. I f the full ram-system flow control i s implemented, the resulting power loss of the ram system can be reduced to approximately the same level as the eductor system. This i s i n direct contrast to the requirements for the flow-augmented eductor system. As shown on the figure, i f flow control i s imposed on the eductor system through a variable-area air intake or some internal device, loss increases as the eductor flow i s decreased. This i s attributed the cruise power to interacting effects of off-design eductor operation (higher pressure loss, incom- plete mixing) being more pronounced on engine performance than the reduction i n ram drag. These performance effects do not include the additional. drags that may be encountered with each of the systems, such as additional wetted area, blockage, and nacelle interference drags with the full ram system.

0 8 3

c d d iated with the propulsion system ~ ~ ~ l 9 a t i o n can result i n a significant percentage of the total effective aircraft drag. ?he specific areas associated with the engine instal lation where the major performance penal ties are encountered should be identified and evaluated for potential improve- ments through improved design criteria.

Fundamental to improving design criteria the definition of a effort thrust and drag accounting method that clearly identifies the interaction of the propulsion system and airframe. These procedures must be defined early i n the preliminary phases of an aircraft program and maintained through flight test.

Through this approach of identification and accounting, a technical data base applicable to each component considered i n assessing the effectiveness of the propulsive effort would be accumulated for defining improved design procedures.

In addition, it would tend to reduce the uncertainties associated with evaluating preliminary aircraft performance.

Specific areas that suggest potential performance improvements on current and future general aviation aircraft are the design considerations used for air-intake sizing on a l l general aviation aircraft, and exhaust duct geometries and cooling system arrangements for propeller-powered aircraft. Studies have indicated that the power loss at typical turboprop aircraft cruise conditions can range from 16 percent (for a stub side exhaust duct, with no flow control installation) to between 2 and 3 percent (for a straight exhaust, ful I flow control system), thus suggesting a 13- to 14- percent improvement i n system performance.

The key t o arriving at a minimum drag, maximum propulsive effort engine installation on any aircraft system i s the interface between the airframe and engine manufacturers. The concept of ''teaming" has been an accepted practice, to a limited degree, among the larger airframe and engine manufacturers for some time.

However, within the last few years, the realization o f the true significance of the concept in terms of achieving the best performing aircraft system (airframe/engine intergration) with minimum cost and program delays has been acknowledged.

From the general aviation point of view, the concept of teaming should be even more significant, since a large percentage of general aviation oiroraft evolve through engine retrofits for performance improvements. I n order to obtain the full aircraft performance potential, the general aviation airframe and engine manufacturer must understand each others sytems i n terms of constraints, performance, penalties, and trade-offs.

The proposed programs for drag reduction are summarized on Figure 6.

v) w K

P 9

w .

K u w k K K 0 .

v) E

e

CJ) E !

n 0 0 0 0 @, North Carolina State University Standard estimation procedures as well as the NCSU "Body" Computer program (Ref, I ) predict that the drag of the two nacelles on the NASA ATLIT airplane will equal the drag of the fuselage. These estimates are based on computations of the drag of isolated nacelle-shaped bodies i n uniform streams with no internal flow. Losses due to air motion through the cooling fins, to helical components i n the flow over the nacelle, or to unusually high levels of streamwise turbulence are not accounted for i n +he analysis, nor are interference effects arising from the presence of the wing or fuselage. The analysis must therefore be regarded as qualitative at best.

Within these limitations, however, one finds that the high drag of the nacelles i s due to their high form drag, this being about three times as large as their skin friction drag. Normally, for a streamlined body the skin friction drag is three times as large as the form drag! When one considers this result and the nacelle shape it seem apparent that the nose of the nacelle i s too blunt, Thus the indicated course i s to increase the dimensions of the nacel e forebody so that the nose (cooling intake) i s relatively less blunt, A prel iminary computer analysis following such an approach (using NCSU "Body") indicates that the reduction in form drag i s much greater than the increase i n skin friction drag which accompanies the increase in surface area. However, to validate this approach it would be necessary to conduct flight tests with modified nacelles during which the total aircraft drag could be determined. Comparison of the drag of the new configuration with that o f the original would then yield the increment (plus o r minus) due to the change in nacelle geometry.

What i s needed, then, is a simple procedure, similur to what i s done in the wind tunnel with clay and wax which permits one to make minor alterations in the nacelle surface contours quickly and inexpensively. A configuration which, I on the basis of computer calculations, seem propitious can then be tested easily.

The results of the tests can then be fed back to correct the estimation procedure.

With such an iterative scheme it seems reasonable that one should not have to test more than three o r four nacelle shapes before finding a practical optimum.

t seems reasona to suppose that these Ile modifications can be effected n polyurethan~ foam t s into the nacel a! wind shear resista foam i s easily shaped and r , it i s conceivable, that fiberglass shells of appropriate contour could be appended to the normal nacelle for the duration of the test. Since the required recontour as l i f t and moment ch necessary to investig modification.

One would probably wish to do so, however, for the configuration finally selected for production, particularly i t s effect on handling qualities at high angles of attack.

The two figures below show the original nacelle as analyzed by the NCSU

"Body" program and an alternate also analyzed by "Body" . These results are the

basis for believing that there i s significant improvement to be gained by recontouring ,he nacelles. If the question of fuel economy becomes critical enowgh, it is to be expected that efforts will also be directed toward treating analytically the internal cooling flow, the propwash components, and the wing and fuselage interference effects .

A completely analytical treatment of the loss in total head experienced by the flow which i s ingested at the front of the nacelle, proceeds over the cooling fins, and then leaves near the rear of the nacelle i s difficult virtually to the point of impossibility. In addition to. the three-dimensional nature of the mu1 tip1icity of tortuous flow passages, one has variable heat fluxes and temperatures at each of these boundaries. The indicated approach therefore is an integral analysis with the magnitudes o f the various contributions to be determined experimentally. By placing a pitot rake and total temperature probe at the cooling air intake a s well as at the cooling air exhaust and measuring the flow areas at these points one can determine the cooling mass flow, its total head loss (which appears as aircraft drag) and i t s heat gain. Comparison of the head loss with that for equal heat addition for flow between parallel plates will then give an indication of how efficient the cooling path is. A significant difference wilt indicate the need for redesign.

Proper analytical treatment of the propwash components and the interference effects must await both the development of accurate, three-dimensional turbulent boundary layer calculation routines and a computer of sufficient size and speed to perform the combined inviscid-viscous problem in a relatively short time, say 30 minutes.

using tufts and/or skin friction t likely to yield tangib References Smetana, F.O.; Summey, D.C.; Smith, N.S.; and Carden, R . K .

"Light Aircraft Lift, Drag, and Moment Prediction - A Review and

Analysis" NASA CR-25 23, May 1975, 492 pp.

n ~xp$oratory Investigation ASA Grant Number NSG 9083 E. J. Cross Mississippi State University An exploratory effort has been undertaken to systematically investigate the drag associated with the cool ing-air flow of contemporary general aviation engine installations. The purpose of this research is to develop a clear specification of cooling drag, provide design data and information, and to develop experimental methods and techniques for determining the value of the cooling drag. It should be noted that this program represents the initial phase of an extensive study o f this subject which will be required in order to develop a full understanding of the physical processes involved. The specific objectives of the program (Figure l a ) are as follows: determine the state-of-the-art which i s manifest by available data and design methods, establ ish appropriate instrumentation and experimental techniques for determining cooling drug by flight test, and determine the relative magnitude and define the significant components of cooling drag. The approach, taken to reach the objectives, is shown in Figure lb.

The flight test vehicle i s a Beechcraft T-34, Mentor, on loan from the Deportment o f the Navy. The T-34, although a relatively old design, is repre- sentative of contemporary, high-performance, single-engine aircraft. The cooling drag will be experimentally determined by two independent methods which will provide a cross-check and the opportunity to correlate techniques.

A complete b;bliography of source material has been compiled that covers the rnid-1920 period to early 1975. Synopses of the available technical papers and reports are being prepared and will be assembled as a compendium of design infor- mation for installation of aircraft piston engines. The state-of-the-art of design factors which are reiewrnt to the general aviation propulsive system installation i s not well represented by pubIications in the open literature. Much of the pertinent data and some of the design methods are proprietary and cannot be obtained for publication, The most highly developed design procedure available in the open I iteraiure i s the Lycoming installation manual which is essentially an adaptation of design methods developed by Pratt and Whitney circa 1945, and specifications of cooling air requirements peculiar to each of the Lycoming engines. Although

Preceding Page Blank

i s ~nconcl~s~ve, i t i s a noteworthy addit~on to the literature and eir exceptional efforts, A work task has been unde~aken to a design manual that wi 1 include inputs from the engine installation engineers and airframe propulsive system designers. This manual w i l I incorporate current design practice to the extent that company propriktary policies will permit release and publication of data and procedures.

A general purpose instrumentation system has been designed and fabricated for the measurement of pressures and temperatures in and around the engine cowl/ nacelle. The system i s modularized and is easily portable and can be moved intact to other test vehicles. It is completely self-contained and does not rely on the host aircraft for power or support. The measurement system i s composed of three synchronized 48 channel scannivalves which provides for 144 pressure data-points, 20 thermocouples for the measurement of engine internal cowl temperatwres, an air-speed transducer and an altitude transducer. All of the data are synchronized to a crystal controlled clock, This system i s shown schematically in Figure 2.

Installofion of two flight t e s t booms incorporating total and static pressure probes, and pitch and yaw servos has been completed. Calibration flighfs have been completed and demonstrated satisfactory performance of each of these. These are self-aligning probes and have virtually no position error from 70 knots through 150 knots. In addition, an outside air temperature probe, and a shielded thermister, These probes are the primary source of h e beeninstalled on the lower left wing.

aircraft performance data, altitude, airspeed, etc.

An array of total pressure and static pressure probes and thermocouples has been installed in each of the inlets and augmentor tubes as shown in Figures 3 and 4 respectively, The engine baffle i s instrumented similarly to standard Lycaming test cell practice (Figure 5). In addition, total and static pressure probes and thermocouples are located at several position in the upper and lower cowl to provide flow dafa throughout the cowl. Surface pressures are being measured at points extending along lines from the inlet iip to the firewall. The cowl i s adequately instrumented to allow calculation of all the pertinent characteristics of the internal flow.

The data are recorded on board the aircraft in analog form on a seven channel FM/FM analog tape recorder (Lockheed Model 417). The data recorded are : 3 channels of scannivalve measured pressures, 1 channel of airspeed data, 1 channel of altitude data, 1 channel of multiplexed temperature data, and a channel of master clock data. The master clock data are used to time synchronize converter that i s use to convert the recorded data into the analog to di agnetic tape, The d i i i magnetic tape inter- IVAC 1106, which i s faces directly with the University mainframe computer, a used for data analysis and manipulation. The data are converted to engineering units and plotted at the computing center. A secondary instrumentation system has been installed on a photo panel to provide a redundant source of aircraft performance data. The panel has a calibrated airspeed indicator, cal ibrated al timeter, clock, outside air temperature read-out and a binary counter. Data are recorded on a 16 m m film at a rate of 1-frame per second.

The flight test program consists of six schedules which involve calibration of the Pitot-static system, calibration of the primary instrumentation system, gliding flight drag polars for three cowl configurations, and cowl performance with engine power. All calibration flights and approximately 80% of the gliding flight schedules have been completed. The flight test procedure for developing drag polars consists essential I y of a series of s a w tooth climbs and power-off glides at constant cat ibrated airspeed. A drag polar i s generated for each of the aircraft test configurations as illustrated in Figure 6. I n addition, cowl internal flow data is accumulated for each flight condition 50 that momentum changes of the cooling-air flow through the cowl can be compared with changes in total airplane drag indicated by drag polar shifts.

All glides are with the engine off and propeller feathered. The propeller was obtained on loan from HartzeII Propeller and the governor and unfeathering accum- ulator from Woodward Governors. This system provides increased safety and flex- ibility during the power-off gliding flight tests.

The three cowl configurations are: the standard T-34 arrangement, inlets blocked SO there is no internal flow, and the augmentor tubes fixed with butterfly valves in each to throttle the cowl flow. Changes in total airplane drag, changes in the momentum of the internal flow, and changes in the external cowl pressure field are determined as a function of flight condition and air mass flow rate through the cowl. The drag associated with the engine installation and the internal flow of cooling air i s determined by comparing the airplane total drag for each cowl configuration to the case of no cooling air flow with the cowl closed. Airplane drag increments due i o changes in airframe parasife drag caused by perturbations to the external flow induced by inlet spillage are also attributed to cooling drag.

A steering committee has been established to provide a working interface between potential industry users and the University research team. Industry members are from each of the following companies: Avco-Lycoming, Teledyne-Continental, eech air craft^ Cessna ~ i r c r a f t ~ iper A~rcraft Forma ice each year at the Universi requent visits are planned at the individual company f a c ~ l ~ ~ ~ e s . The purpose is to establish a mechanism for the exchange of ideas between the research group and the industry design group to insure the validity of program objectives and increase the probability of useful results and the direct transfer of technology developments to appl ication, The NASA Technical Officer is Mr, Albert W. Hall Mail Stop 247, RAFD NASA-Langley Research Center Harnpton, Virginia 23365.

a : !ii

E"

u , I - cn W I -

.-

U n W Y

.- 5

.I- -I l o o .

m

.- s

c .I- S aa L + n c w h a L

i k

if L .

m rl .

L .

N m - 4 I a aJ 0 a a d v) k 1 . I 'd \o aJ U :: 4 0 c u a c H

i

W a , r :

.-

I

.

U v1 aJ H U %I rl d r4

e

E. E . Larrabee Massachusetts Institute of Technology Abstract The fundamental vortex theory for a single rotation propeller with a finite number of blades i s reviewed. The theory leads to the specification of a radial distribution of bound circulation on each blade for minimum induced loss, analogous to the elliptic spanwise distribution of bound circulation on a wing for minimum induced drag. A propeller designed in accord with this theory has been tested in the water tunnel at M.I.T.'s Marine Hydrodynamics Laboratory where it exhibited high efficiency in spite o f localized cavitating flow. A knowledge of the flow fieid for an optimum propeller i s of value to the airframe designer seeking to maximize the performance of the airpiane-propel ler combination.

Figure 1 presents the geometry of the force and velocity components asso- ciated with the operation of a representative blade element. The fluid velocity W at the blade element i s made up of the flight, or advance velocity, V, the rotational velocity, Rr, and the induced, or "inflow" velocity wi. The induced velocity i s customarily resolved into an axial component aV and a rotational component a%r. The elementary force d F i s resolved into lift and drag components dL and dD where the angle f , defined as tan-' (cdck>, is an angle determined by blade profile characteristics in two dimensional flow at the appropriate Reynolds number, as in lifting line theory.

The blade element efficiency i s given by - - .- and since while it follows that tan 4i

"element - - tan ($i+c) l+a - ' p r o f i l e 'induced

l-at -

X In this discussion i w i l l be mainly concerned with the "induced" efficiency (1 -a' )/(I*).

Figure 2 presents the geometry of the velocity field i n the propeller sl ip- stream. An elementary helical vortex filament i s convected normal to itself with velocity w by the induced velocities of a l l the vortex filaments lying in the approximately helicoidal vortex sheets trailed by each of the propeller blades.

The vortex velocity w, which i s the same as the local slipstream velocity, may be resolved into axial and rotational (or "tangential") velocity components wq and wt, respectively.

If the filament helix angle i s 9, the filament will appear to move with an axial velocity v', which might be due either to real axial velocity wa or i o real rotational velocity wt (like the rotating stripes on a barber pole); but since it is a vortex filament, and can only move normal to itself, . .

wa = v ' cos $, and wt = v' cos 4 s i n 4 Figure 3 presents the so called "lightly loaded" propeller relations between velocities wq and wt in the slipstream at radius r and the corresponding inflow velocity components aV and a'Rr at the propeller disc. Since the propeller i s lightly loaded, the trailing vortex helix angle 4 in the slipstream at radius r is indistinguishable from the angle tan-' (V/Q r) at the propeller blade element at the same radius. Following momentum theory the inflow velocities are taken as half their values i n the developed slipstream: ---- The induced efficiency of a blade element may thus be expressed i n terms of the apparent axial velocity of the slipstream, v', which up to now has been considered to be an arbitrary function of r. Betz showed induced for the propeller as a whole i s maximized i f nelement i n 1919 that i s the same for all blade elements, that is, if v'/V i s independent of r . I f v' i s independent of r, all vortex filaments i n a helicoidal vortex sheet appear to move axially as a rigid surface, although this i s not actually the case, since wq and wt are given by eqs. 5 and 6.

Since circulation cannot be added to the flow in the slipstream, it follows that the circulation within a slipstream tube of radius r

r = 2 7 ~ ~ w 10

t must comprise the total circulation of the helical vortex filaments trailed by each of the B blades of the propeller, an amount equal to the total bound vorticity at the radius r. The bound vartjcity on each blade at radius r is then or 1 la where x 2 Slr/v, Since, B, SI, 2v, V, and v' are all constants for a minimum induced l o s s 2 2 propeller, B.Qr(r)/ZTVv' = x /(x + l ) , can be regarded as a normalized form for the bound circulation, T (r), expressed as a function of the normalized radial coordinate, 2 2 2 x = Qr/V.

Alternatively, the quantity x /(x + l ) , which i s also equal to cos 4, may be regarded as the ratio o f the axial velocity in the slipstream to the apparent velocity, wa/v'. Figure 4 presents the normalized radial circulation distribution (or the w d v ' ratio) as a function of the normalized radial coordinate, x. The ratio of the rotational velocity in the slipstream to the apparent velocity wt/v', equal to x/(x -+I), i s given for comparison. It i s seen that single rotation pro- pellers of minimum induced loss inevitably have appreciable slipstream swirl near s large; the swirl angle i n airplane -1 coordinates being give y by tan [w+/(V+wa)l.

a apply to a propeller with so many blades Strictly speaki i s s m a l l that the spacing between the individual vortex sheets i n the slipstream compared to r, called the B-+w case. Since actual propellers have a small number of blades and may operate at large helix angles, which also tends to increase the spacing between vortex sheets (note that rlprofi le = tan $i/tan (@;+E) i s maximized

when $ i = 7r/4 - e/z), it i s necessary to account for the reduction in the average

rotational velocity between vortex sheets in the developed SI ipstream {compared to the v’ cos + sin @ value at the sheets themselves) when calculating the circulation.

Figure 5 shows Prandtl’s approximate solution to this problem. He assumed that the flow near the edges of the helicoidal vortex sheets in the slipstream i s like the two dimensional flow near the edges of a semi-infinite array of flat plates moving with velocity v. The average velocity of the fluid between the plates i s given by the fraction F = ( 2 / ~ )cos-’ e-f times v, where f = ~ ( ( / s ) i s a dimension- less measure of the distance € from the edge o f the plates spaced a distance s apart.

The corresponding edge distance function for the helicoidal vortex sheets i s where h = V / S 2 R is an advance ratio based on the flight velocity V and the rota- tional tip speed SIR. The quantity F is interpreted as the ratio of the average rotational velocity in the slipstream at radius r to the rotational velocity nearthe is the same thing, the ratio of the bound circulation sheets, v’sin@cos$, or, what r for a propeller with a finite number of blades to the corresponding at radius circulation for B + m * Figure 6 presents representative examples of the radial distribution of bound loss three blade propeller operating at two circulation for a minimum induced advance ratios, X = 1/5 and X = 2 / 3 , corresponding to climbing and cruising flight conditions, respectively, where F has been calculated according to Prandtl‘s rule. These optimum radial circulation distributions for propeller blades loaded so as to produce constant apparent velocity of the helicoidal vortex sheets in the SI ipstream, independent o f the radius, correspond to the el liptic spanwise distribu- tion of circulation for a wing, loaded so as to produce constant downwash velocity at the trailing vortex sheet in the wing wake, independent of a spanwise coordinate.

Although an untwisted wing of elliptic planform gives an elliptic spanwise c~rculation d~stribution for all angles of attack within t e linear range of section lift coefficient versus angle of attack, there i s no correspon plan for^ that gives the optimum radial circulation d ~ s t r ~ ~ u t ~ o n for a l l advance ratios. Figure 8 presents plots of (Q/Tv')cc, versus Ax = r/R for three bladed propellers o f minimum induced loss operating at the two advance ratios given on Figure 7. If the section lift coefficient, cE, i s considered constant for a l l values of r/R, the curves may be interpreted as plots of the optimum ra I chord distribu- tion. It is seen that a propeller optimized for a low advance ratio needs blades with a wide chord inboard where the airspeed i s low, and a narrow chord outboard where the airspeed i s high. A propeller optimized for !arge advance ratios, on the other hand, will require a more elliptic distribution of chord since the blade element air- speeds are not s o strongly dependent on the radial coordinate. I n any event, i t i s seen that a propeller can only be optimized at one advance ratio and that a general theory of non-optimum propellers will be required to calculate propeller performance away from the design point.

A traditional method for doing this was developed by Glauert. In his scheme the differential increases in axial and rotational momentum in an annular element of slipstream of radius r and width dr are set equal separately to the thrust and toque components of the airload acting on each of the B blade elements, thus: 1 2

dT = 2 n p r ~ ( l + a ) 2 a ~ = --pw B ~ C

dr 2 Y

and where

C = C ~ C O S + ~ - cdsin$

Y i

Cx = cRsinGi + C ~ C O S ~ $ ~

Equations 13 and 14 are adapted for calculation by solving 13 for a and 14 for a', where a i s the local solidity, Bc/&r: ( J c a 13a L sin +i n iterative procedure i s employed at each of several blade radial stations y blade element angles o f crttack are assumed giving CR and cd ( = @-a ( B is t h e blade a n g l e ) , ' i c and Cx (eqs, 7 5 and 161, Y a and a' (eqs. 13a and 14a), and finally -1 V l+a +i = t a n (E) ( 1 4 for each assumed value of a . The iteration i s terminated when +i = B-z= tan-'(&)(E') and the converged values of C and C are then suitably integrated Y X radially to yield the propeller thrust and torque (or power).

Equations 13 and 14 depend on the absence of radial flow in the slipstream, the very thing that Randtl's vortex sheet spacing correction was intended to account for in the case of a propeller operating with minimum induced loss. Glauert i n his article i n Durand's "Aerodynamic Theory" (1934) suggested modifying Eqs. 13a and 14a to read 13b 14b (there i s a misprint in the book whereby the quantity F appears in the numerator of the right hand side of eqns. 13b and 14b instead of the denominator). Goldstein, in his doctor's thesis (published as " O n the Vortex Theory of Screw Propellers" in the Proceedings of the Royal Society (A) 123, 440 1929) refined Prandtl's value for F by considering the flow about moving helicoidal surfaces of B sheets per turn rather than an array of moving flat plates. F. N. Lock proposed an alternative

scheme based on Goldstein's values for F (Lock calls them 3, sometimes read

"kappa") in which the momentum balance of eqns 13 and 14 i s abandoned and the inflow velocity, wi, (Fig. 1) i s considered to be normal to the resultant velocity at the blade element, W. The inflow velocity, i n turn, i s considered to be half the developed Slipstream velocity increment, w, given by a form of eq. 10: ock's pro~edure w o u i ~ be ~dent~cal to lauert's modified sc eme if the blade elements had no dra e Theodorsen introduced a correction to Lock's pro allow for slipstream contract~on; he also verified Goldstein's F values by rheo- electric analog computation, and extended them to the case of intersecting heli- coidal vortex sheets, as would be trailed by a counterrotating propeller. Lock's method has recently been reviewed by Pauiing ("The Effects of Unce Predicting Rotor and Propeller Performance", Pennsylvania State Un PSU AERSP 75-3) who wrote a digital computer program to carry out a version of Lock's procedure with several optional features, including a slipstream contraction effect. I personally am bothered by the fact that all of these non-optimum propeller theories depend on Prandtl or Goldstein F values, which are calculated on the assumption of trailing vortex sheet geometry appropriate to a propeller of minimum induced loss. The procedures are analogous to an approximate lifting line wing theory in which the two dimensional lift on a chordwise element i s corrected by [l-(2y/b)2J1b to account for three dimensional flow effects.

Figure 9 presents a marvelous smoke flow visualization photograph of the operation of a two bladed propeller obtained by Prof F. N. Brown at Notre Dame.

The helicoidal vortex sheets are seen to roll up rather rapidly as they are left behind in the slipstream by the propeller blades, exactly i n the way that the trailing vortex sheet left behind a wing does. The picture suggests that the propeller i s not optimally loaded because there i s perhaps not enough axial motion of the inboard portions of the trailing vortex sheets--although this i s difficult to judge, because

the sheet i s marked by smoke particles which have - both rotational as well as axial

velocities, while the "apparent" v' of the theory specifies the motion of the heli- coidal past a fixed point; for example, the "apparent" axial velocity v' component due to "barber pole" helix rotation, Vlbp = wttan$, does not show in the picture.

I will conclude this presentation of propeller theory with the observation that the propeller equivalent of lifting line theory does not exist, and that a l l propeller computation procedures contain some element of empiricism. Helicopter aerodynamicists have had some success in the application of machine based discrete vortex models to the prediction of rotor characteristics; see for example Landgrebe's "The Wake Geometry of a Hovering Helicopter Rotor and i t s Influence on Rotor Performance" (Journal of the American Helicopter Society, Vol 17, no. 4, October 1972), but in my opinion work s t i l l needs to be done.

0 shows a resea h propeller constructed at e water tunnel of the ydrodynamic ~ a ~ ~ a t o ~ e Th to have two blades of a ~ i n ~ m u ~ induced loss geometry appropriate in an app~~cat~on to CI direct drive 1700cc Volkswagen engine installation which develops 47 hp at 3800 rpm at sea level The full scale propeller would have a diameter of 50 inches and be optimized at 120 mph, giving an advance ratio J = V/nD = 0.667 (A= 0.212). The model propeller had a diameter of 12 inches and was constructed so it could be tested either in a four blade or a two blade version, since additional blades could be readily made once the milling machine cam had been constructed to make one blade.

Figure 11 shows the propeller operating near i t s design advance ratio i n the water tunnel at a pressure low enough to cause appreciable cavitation over the outer quarter of the blades. Note that the propeller i s tested in a pusher configura- tion with the shaft extending upstream into the tunnel stilling section, and with a spinner fitted downstream to preserve good flow at the inboard blade stations. The compound helical character of the cavitation marked tip vortex core i s noteworthy.

Figure 12, finally, presents the measured characteristics of this propeller as tested in the water tunnel at a loading high enough to produce light cavitation.

in the upward bulge and the downward dip of the The effects of cavitation are seen torque and thrust curves, respectively. The peak efficiency of 85%, obtained at an advance ratio, J = 0.8, is the highest ever measured in this tunnel.

It is hoped that this paper will encourage general aviation aerodynamicists to seek propeller geometries better suited to the operation of their own airplanes than the compromise production propellers available as off-the-shelf items. It should be borne i n mind that the interference flows produced at the propeller by a large fuselage downstream need to be taken into account in the design of an actual propeller, and that efficient propellers inevitably create large slipstream swirl components on the fuselage nose and flanks which should be considered in the design of engine air inlets, carburetor air scoops, exhaust stacks, landing gear struts, and even door handles.

d

E

b .

3.

N

!h

”;..

If .

.x .

sc

-8

c;

+

R,

L

I t !

c - ~ ~ - i i I I i .

a

. .

. .

! I .

. . .

1 1

. 1 I

t

i .

V - T G 7 .

J , Roskam, University of Kansas 7.2 Trim Drag Research Results H. Chevalier, Texas A & M University Reduction of Trim Drag in General Aviation Airplanes 7.3 F. H. Luke, Virginia Polytechnic Institute Trim Drag in the Light of Munks Stagger Theorem 7.4 E. E. Larrabee, Massachusetts Institute of Technology

ing Page Blank

e Comments on Trim b a g Jan Roskam University of Kansas Introduction This paper presents a discussion of data of and methods for predicting trim drag. Specifically the following subjects are discussed: - Economic impact of trim drag.

- The trim drag problem in propeller driven airplanes and the effect of

propeller and nacelle location.

- T heoret ica I procedures for predicting trim drag.

- Research needs in the area of trim drag.

An Example of the Economic importance of Trim Drag Trim drag is here defined as the horizontal t a i l induced drag caused by the need to trim the airplane for Cm = 0. Tail profile drag i s included in overall airplane zero I if t drag.

Trim drag typically varies from .5 percent to five percent of total airplane drag in cruise, depending on airplane type and on center of gravity location.

with center For a typical business jet, Figure 1 shows the variation of AC b r i m of gravity location. Using this example, assuming a cruise L/D = 10.8, a cruise thrust required of 1092 lbs. at M = .72 and.45,000 ft., Table 1 shows the fuef flow caused by this drag for three c.g. locations.

Table 2 summarizes what this means for a n operation using o n e airplane 1000 hours per year. Table 2 also shows what the fuel expenditures due to trim drag are for a fleet of 500 airplanes in one year.

Although trim drag by itself seems so small a s to be negligible, integrating it over time and fleets indicates that more careful attention should be paid to trim drag.

Preceding Page Blank

! i . .

e . . . . . .

. .

-

5 C.G. LOCAT\Or3 ohf c

Figure 1 . Example of Trim Drag Variation with Center of Gravity Location

Table 1 . Fuel Flow due to Trim Drag a s Function

of Center of Gravity Location . _ ' , ~ ._-_-- _. ._. - . .. -- .-- I Fuel Flow due t o T r i m Drag Center o f Gravity { 1 bs. /hr. ) (See Fig. 1) 19.5 M I D

-_ ..-

AFT LIMIT Table 2. Economic Importance of Trim Drag .

A f t C.G. Fwd C.G.

extra gal 1 ons extra gallons burned due t o burned due t o t r i m drag t r i m drag _. - -.

1000 hours each 91 6,000 2,977,000

- It would seem that the designer, when trying to

many unsolved problems. To illustrate the complexity of the design problem when including trim drag, consider Figure 2 and the following equations which need to be satisfied:

c , = e w e +

It is noted thut a l l coefficients and derivatives in equations 1 through 5 are functions of the shape of the configuration (including fuselage camber) and the location an angular orientation of the thrustline. The question which needs to be answered is how to optimize L/D. In view of potential importance of trim drag and of associated design decisions' with *e hand1ing qualities of the the interaction this area seems needed.

airplane, some theoretical (methodological) research into Certainly no solution to this problem is readily available today, except perhaps in the case of pure jet u irplanes .

trim drag i s In nearly all current propeller driven general aviation airplanes, ignored, so that the entire problem of trying to minimize it a s part of the overall drag does not come up.

Illustrations of the Effect of Nacelle and Propeller Location on Trim Drag -

Reference 1 shows the importance of thrust coefficient on C b of different airplane configurations. Figure 3 illustrates the favorable effect increasing thrust coefficient

can have on CM . At the same time, Figure 4 shows how decreasing wing height

Figure 2. Illustration of Design Choices Affecting Trim Drag

can have an unfavorable effect on CMo. (A change in C of I. 10 1 , using

M O CM& E = -.02 means a 5 degree change in elevator required for trim.)

Reference 2 shows that a downward t i l t of 5 ’ of the propeller axis of a typical W I I fighter configuration can cause an aft shift in a.c. of 5 to 10 percent, while also causing very large changes in C M ~ . Even though the aft shift in a.c.

may be desirable to attain satisfactory longitudinal stability on high horsepower configurations the effect on trim drag is unfavorable.

These illustrations are meant to show the importance of considering the complicated interactions of these factors. N o simple, reasonably accurate preliminary design procedures exist to account for them. Evidently there i s a need to develop them.

A Method for Predicting Trim Dran of Jet Airdanes Figure 5 illustrates the relation between WBV-lift and H-lift vectors. Note that it i s not immediately clear from Figure 5 whether overall lift-to-drag ratio improves or deteriorates with c.g. movement. This depends to a large extent on the of WBV i n its untrimmed state relative to the value of L/Dlmax of WBV. It evidently also depends on whether the tail i s uplifting or downlifting to achieve pitching moment balance. (These comments also’apply to propeller driven airplanes .)

Figure 6 illustrates the possibilities and Figure 7 shows the potential outcome.

Reference 3 shows that the trimmed drag coefficient of a tail-aft configuration can be estimated from: Figure 3. Effect of Propeller Operation on the Pitching-Moment Coefficient -4 0 4 8 12 16 Anpk'of onock, d .&g Figure 4. Effect of Propeller Operation on the Pitching-Moment Coefficient W l l M LWAC t SUIICICL A t R ? U N € WW4N TO HIGWLIGMTSGN NOTk CONVLNTIOI( FOIl AFT TAILS A N 0 CANAIIOT. C J I A L W TRlAlS a WRtACt AIRPLANES ONLT - .

Figure 5. Coordinate System and Sign Convention w ZsROLOAO R F E 3 (c) LlFTlWO APT (c) LlFTlWO APT TAIL TAIL Figure 6. Effect of Center of Gravity and Wing-Bod Aerodynamic Center Location on Lift Sharing C M ~ = J.

--- - TAIL PROFILE DRAG A N 0 I N ~ L I N A T I O N DRAG

..*...*...- TAIL DRAG DU6 TO LIFT

0 ZERO LOAO TAIL DOWN LOAO TAIL LIFTING APT TAIL OR CANARD

OPEN SYMBOL - WING-BODY LIFT AND DRAG

SOLI0 SYMBOL- TOTAL TRIMMED LIFT A N 0 DRAG LIFT Dww Figure 7. Trim Drag Elements C D T R t M O W B V Note the absence of thrust effect terms.

Expressions for C L ~ ~ V / C L ~ ~ ~ ~ and C L , / C L ~ ~ ~ ~ can be found by imposing the conditions that total lift equals airplane weight (level flight) and that the total pitching moment i s zero: -_

+ C c,= Lo

wev

w"w%v cx + c,, (3 1

where

culw= - C ' , ' H - %H

c z 1

?

and Reference 3 shows for fighter type configurations that this method gives accurate results. Results indicate that trim drag can affect the trimmed L/D of such configura- tions by 2 7 percent depending on overall arrangement and c .g. location.

From equation (7) it i s evident that CM can play Q role in reducing adverse (i.e., down load on tail) trim requirements. It would be desirable to give the air- plane a positive value of CM > 0 by fuselage camber. What is not known today, is how the general aviation'fuselage can be shaped in such a way that: 1 . C b i s as close as possible to being positive 2. f w d visibility and windshield shape are compatible with C b > 0 and low windshield drag contour 1 ines are not expensive to produce 3.

aft fuselage shape does not violate take-off rotation requirements.

4.

- - - Some systematic research into this area may very well pay off. Perhaps a theoretical trade-off study of a wide range of fuselage camber shapes should precede a systematic windtunnel investigation.

The effect of wing mounted nacelles on C and Coo should also be Mo investigated in a systematic manner. The latter i n view of the fact that general aviation twins use widely varying wing-nacelle shapes not all of which can be particularly good. (See Reference 4.)

mary of Research n view of the fact that t r ~ m drag can affect the cruise ift-to-drag drag ratio y up to seven percent, it WQU desirable to have procedures available to accurately account for it. For propeller driven airplanes these procedures do not seem to exist.

Because of the potentially large effect of C E J ~ on trim drag, this quantity should be accurately predictable. It is not today.

The following research is therefore needed: Development of a theoretical procedure to predict C b including 1.

propeller thrust interactions; Development of a preliminary design method for predicting trim drag of 2.

propeller configurations; and Configuration research to see if perhaps other than conventional 3 .

tail-aft configurations are capable of yielding better cruise I ift-to-drag ratios References Katzoff, S . ; Longitudinal Stability and Control with Special Reference to 1 .

S I ipstream Effects; NACA TR 690, 1940.

2. Goett, H.J. and Delany, N-K.; Effect of T i l t of the Propeller A x i s on the Longitudinal Stability Characteristics of Single-Engine Airplanes; NACA TR774, 1944..

Goldstein, S.E. and Combs, C.P.; Trimmed Drag and Maximum Flight 3 .

Efficiency of Aft Tail and Canard Configurations; AIAA Paper 74-69 presented at the AIAA 12th Aerospace Sciences Meeting, Jan. 30 - Feb. 1 , 1974, Washington, D .C.

Roskam, J .; Drag of the Complete Configuration, Part 11, Aerodynamic 4.

Considerations; Paper presented at the NASA/Industry/University General Aviation Drag Reduction Workshop; July 14, 15, 16, 1975, The Univenity of Kansas.

Texas A & M University This paper was not submitted f o r inclusion in these proceedings.

Preceding Page Blank

~ ~ e c ~ n ~ c Institute and State Liiversity It i s important at the outset to distinguish between "trim drag" and "trimmed drag. 'I According to the USAF Stability and Control Handbook,(') the trim drag coefficient i s "the drag coefficient increment between the drag coeffic complete vehicle in pitch equilibrium and the drag coefficient of the wing-body- vertical tail Configuration. " The trimmed drag coefficient, on the otherhand, i s the drag coefficient of the complete vehicle in pitch equilibrium. It is clear that our interest should be focused on reducing the trimmed drag and not on the nebulous problem of reducing the trim drag penalty. Consequently, emphasis will be placed on the complete configuration and the associated trimmed lift and drag with particular attention paid to the load distribution between the wing-body and the tail surfaces.

Aircraft Equations for Equilibrium, Balance, and Drag The equations for the total aircraft lift and pitching moment coefficient are given by (for small downwash, c ) (21, (6) and For balance in equilibrium flight, C, = 0, allowing equations ( I ) and (2) to be solved for the tail lift coefficient and the aircraft angle of attack:

Preceding Page Blank

t cL c

*- -

C mowb

a -a I -

(4) owb Equations ( 3 ) and (4) govern the distribution of the required lift force between the wing and the tail and insure a zero pitching moment. Several observations can be made concerning these equations: The tail contribution to the aircraft lift coefficient, (i) Vt St/S, i s a function of wing-body properties, c.g.

( C , ) = C

Lt

posittion, and lift coefficient. Consequently for a given speed

and weight, (CL), , the tail load can be adjusted by shifting

the c.g. position or by changing the wing-body aerodynamic characteristics.

(ii) The expression in the denominator common to both equations is independent of the c.9. position.

(iii) The magnitude of C L ~ determined by equation (3) must be less than CLtmax* The key to the selection of wing-body parameters and c.g. position is the introduction of the alrcraft drag coefficient. We would like to select these para- meters to reduce or minimize the drag coefficient for a given lift coefficient. The the aircraft i s given by (for small downwash c): (6) drag coefficient for . - where The bracketed term in equation (5) i s the trim drag coefficients as indicated by the definition at the beginning of the p o p .

The problem of the aircraft designer then is to select the vving-body aero- dynamic parameters and c.g. position such that the trimmed drag coefficient given uat~on (5) i s m i n i n some sense, subjected to the e straints given by equations (3) and ( 4 ) . Co nt s t a b ~ l ~ ~ spec~~cations requ~re that ce Consequently i n current design practices the stability and performance characteristics of an aircraft are virtually determined independent of each other in the sense that one aspect (stability) i s considered and then the other (performance). (3) The continuous improvement of digital computational equipment with respect to size, speed and reliability has led to the consideration of utilizing digital control systems to maintain stability, reducing the number of constraints on the selection of c .g. position and wing-body aerodynamic parameters to reduce the aircraft drag

coefficient . (4)r These increased degrees of freedom present a considerable

challenge to the aircraft designer leading to some of the new concepts of design associated with controlled configured vehicles. Although it i s anticipated that such sophisticated control systems will not be available for general aircraft for a considerable period of time the advantages of such systems should not be completely ignored.

In whot follows the concept of reducing the aircraft drag coefficient by appropriate selection of the wing-body aerodynamic parameters and c . g . position, w i l l be examined. This approach i s equivalent to finding the minimum drag for a given speed as opposed to muximum L/D for the aircraft.

C. G. Position for Minimum Drag If we ignore stability requirements it is possible to determine the c,g.

position which minimizes the drag coefficient for a given lift coefficient. In order h accomplish this, the appropriate terms in (5) are replaced by the expressions given in (3), (4) and (6). Furthermore the c.g. position can be introduced by noting the following relations: where h, i s position of x in chord lengths behind the leading edge of wing and x = 0 c.g. position = t tai1 aerodynamic center = nwb wing-body aerodynamic center t with respect to c .g. location can be eva~uated and set e resulting expression can then be so ovides m ~ n i ~ u m drag coefficients. The result i s h, =

2[a (k +k') - .%

w b wb t a , ] 'L I where kt = kt/(ntSt/S) Equation (8) gives the c.g. position for given lift coefficient for minimum drag coefficient in terms of wing-body aerodynamic parameters, tail parameters and geometry.

Several observations concerning equation (8) can be made: (i) The c .g. position for minimum drag coefficient changes with lift coefficient (speed) (ii) The c .g. position dictated by (8) i s not restricted by stability constraints allowing the possibility of inherent static stability (iii) The c.g. location given by (8) i s a function of wing-body and fail aerodynamic parameters and geometry. Consequently the c.g. location for minimum drag coefficient can be changed by judicious selection of these parameters.

Design Characteristics As indicated earlier it i s undesirable to have an inherently unstable (or overly stable) aircraft when sophisticated control systems are not available for compensation purposes. Consequently it would be desirable to take advantage of observation (iii) and adjust the aerodynamic and geometric parameters in such a manner so that the c.g. position for minimum drag provides the desired static margin. It i s possible to approach this problem several ways, two of which will be out1ined below.

One method of approach i s to treat the drag coefficient as a function of several aerodynamic and geometric parameters, includf g c.g. position and attempt to find a minimum with respect to all these parameters subject to certain specified constraints (static margin, etc.). The drawback with such a method i s that a large 31 0 led to obta~n "optimal " pa~~meters which ach takes advantage of equation (8) and the related observa- (iii). Here the optimal c.g. position a s a function the aerodynamic and geo- tion metric parameters i s determined by (8). Furthermore, the drag coefficient and the neutral point position can be determined in terms of the same set of parameters.

For small changes in the parameters we can approximate the changes i n drag coefficient, c,g. position for minimum drag coefficient, and neutral point by: Consequently for a given aircraft, the "optimal" c.g. position can be selected from equation (8). Then equations (9) can be used to find the changes in the para-

meters pi required to move the c . g. and neutral point to satisfy static tnargin

requirements and at the same time keep h C ~ k 0. I n other words A h , , A h , , and ACD are specified and (9) solved for pi. If there are more or less than three para- meters the solution i s either nonunique or not possible. In such a case a minimum norm. type solution is proposed.

The changes in the parameters can be incorporated by appropriate changes in the wing-body and tail geometry, another area which needs development. Again several observations can be made: Clearly a necessary assumption i s that the parameters can be (i) changed independently. This assumption is better for small changes in parameters and decreases in i t s validity a s the magnitude of the changes increases. ' The calculation of sensitivities in the above method allows an (ii) evaluation of the importance of each parameter in achieving a desired goal.

3 7 1 ed at exam~ning methods for reducing the drag coefficient for a given aircraft lift coefficient, or speed. The emphasis was placed in determining the load distribution between the wing-body combination and the tail which would reduce the overall drag coefficient. Furthermore a technique was presen d which would allow the determination of var dynamic and geomet parameters which would permit the 'best' c.

satisfy inherent stability requirements. Included in the method was the calculation of sensitivity coefficients which indicates the importance of various parameters in achieving specified goals ie. c .g . movement, drag coefficient change, etc.

Preliminary results indicate that such an approach i s feasible. For given aircraft parameters c.9. movement alone yields drag coefficient reduction of the order of 1% over the nominal case for a conventionally designed aircraft. Tentative results indicate that if the downwash angle at the tail i s large enough (at zero lift) then a down load on the tail at the expense of the same additional load on the wing is desireable in reducing the overall drag coefficient. The reason i s that the t a i l If the tail lift is negative the lift vector i s tilted rearward by the downwash angle.

.

contribution to the aircraft drag is negative. Under these circumstances the optimum c .g. i s forward of the nominal. The amount and direction o f movement i s sensitive to this downwash parameter.

Although the drag reduction due to c.9. movement alone i s small, the inclusion of other parameter changes can improve this drag reduction significantly.

How these desired parameters changes can be obtained through wing-body and tail geometry changes still needs to be investigated. Clearly all drag reduction methods should be examined together. (7) York, 1972.

3. Goldstein, S. E. and Combs, C. P., "Trimmed Drag and Maximum Flight Efficiency of Aft Tail and Canard Configurations," AIAA Paper 74-69, 12th Aerospace Sciences Meeting, Washington, D. C. 1974.

4. Luke, F. H . and Cliff, E. M., "Control-Configured General Aviation

Aircraft, SAE paper 730303, Business Aircraft Meeting, Wichita, Kans.

April 1973.

5. Hood, R.V., "Active Controls Changing the Rules of Structural Design," Astronautics and Aeronautics, August 1972, pp. 50-55.

6. Hofmann, L. G. and Clement, W.F., "Vehicle Design Considerations for Active Control Application to Subsonic Transport Aircraft," NASA CR-2.

August 1974.

7. Rediess, H.A. (Editor) "Advanced Control Technology and i t s Potential for Future Transport Aircraft, I' Preprint from NASA Symposium of same name, July, 1974.

31 3 Load Distribution for Various CG Positions (Speed Changing ) 31 4 Load Distribution for Various Speeds (CG Changing) CG Position for Minimum Drag Coefficient 31 6 Induced Drag vs CG Position 31 7 .

induced Drag vs CG Position 31 8 E . E, Larrabee Massachusetts Institute of Technology Abstract Munk's stagger theorem holds that the induced drag of a multipla dependent of the streamwise position (the stagger) of its lifting elements so long as the gap/span ratios and the element/element lift ratios are specified. In particular, a monopiane-tailplane or a monoplane-foreplane (canard) arrangement can be re- garded as a biplane of zero gap and the trim drag due to tailplane download or foreplane uplpad can be readily calculated. The trim drag penalty i s the same for both configurations. Relations are given for trim drag estimates for various practical arrangements.

Max Munk was one of the f i r s t generation of Goettingen aerodynamicists.

old NACA, and was largely responsible for the concept of, Later he worked for the and the first test programs carried out in, the Variable Density wind tunnel. He contributed greatly to our present understanding of aerodynamic drag. While still at Goettingen he discovered some general laws about the induced drag of multi- planes, one of whlch i s set forth in Figure 1.

* Prandti used this law as one of the cornerstones of a monograph on the "Induced Drag of Multiplanes", which appears in German in the Technical Reports (Technische Berichte) Vol. 1 1 1 , No. 7, pp. 309-315 of the aerodynamics research establishment at Goettingen This report was immediately translated into English and published by the NACA as Technical Note No. 182 in 1924. I t s contents also appear i n Glauert's "Elements of Aerofoil and Airscrew Theory" Figure 2 gives Prandtl's formula for the induced drag of a biplane. It i s written as the sum of the self-induced drag of the elements of the biplane, plus twice the induced drag of one element due to the flow about the other for the case of an unstaggered array. Munk showed that the cross induced drags of the two elements were equal for an unstaggered biplane, but that also, by virtue of his

stagger theorem, the - sum of the cross induced drags was unchanged by stagger so

long as the lift distribution between the elements i s preserved. Thus the cross in- duced drag of the forward element of a biplane i s reduced by the upwash about it 31 9 nt; conversely nwash about it due to the itude of the cross induced dra he biplane inter- ference factor,. CT, defined by the relation on Figure 2. I t s numerical value was calculated by Pohlhausen, who graphically evaluated an integral which gave the cross induced drag of one element of an unstaggered biplane carrying an elliptic span loading in the presence of the other element, also assumed to be elliptically loaded, and creating the downwash field appropriate to an elliptic span loading at the arbitrary location of the first element. The values of cf were evaluated for three discrete element span ratios and several gap to average span ratios. The it might be worthwhile to refine Pohlhausen's results are presented in Figure 3.

calculations with a modern calculating machine. For our purposes it will suffice to note that (r approaches the span ratio in the l i m i t as the gap/span ratio approaches zero.

Professor Ober of M.I.T., now emeritus, always taught his students (and I teach mine) that the induced drag of a monoplane-tailplane combination can be closely estimated by treating it as a staggered biplane of zero gap. Figure 4 presents some results of such a calculation. It is seen that the induced drag penalty for carrying a download of 1oo/o of the total lift on the tailplane of a conventional airplane i s slightly more than 1oo/o of the minimum induced drag of the wing alone for a tailplane to wing span ratio of 0.3; surprisingly,. the trim drag penalty for carrying an upload of 1oo/o of the total lift on a canard foreplane of the same span ratio i s identical. The trim penalty disappears for tailplane (or foreplane) span equai to the wing span, as one might expect.

Figure 5 compares the induced trim drag penalties for two representative wing-bodytailplane combinations i n which the tail off pitching moment of the two wing body combinations differ only in the magnitude of the pitching moment about the wing aerodynamic center, the first example corresponding to a conventional NACA 4 digit airfoil, and the other corresponding to a heavily cambered airfoil of the Whitcomb supercritical, or general aviation type. It i s seen that the drag penalties at the rather high total lift coefficient of 0.6 amount to about 0.2 counts and 2.2 counts, respectively; amounts which would be difficult to establish by wind tunnel testing.

Figure 6 presents an experimental verification of this technique for cal- culating the additional induced drag due to tail load by comparison with experi- mental drag data obtained on an 1/8 scale model of the XP-87 airplane during the course of wind tunnel tests conducted to d ~ t e r m ~ ~ e the average dow~wash angle lane as a function of airplane angle of attack with deflected flaps.

The e x ~ e r ~ ~ e n t a ~ tail loads were large and the additional tail drag could be measured accurately, The minimum induced drag of a wing (W) tail (H) configuration may be written as: c c C2 sH 'w LH LH

+ - +-

cD induced sW 'zbwbH/sw) IT &H minumum where + - = b 2 w/Sw AH = b H/SH In this particular case the drag of the complete airplane was computed from the relation -2 L S v - sH Lw

+ - Cd + -

CD = CD +?l v ' W CdH '~BNF where

CL = CL - - sH c

W =H = 0.877 = 0.0660 =WBNF cD '~BNF 'dv = 0.01 Sv/Sw = 0.146 CdH )= 0.01 SH/Sw = 0,219 bH/bw = 0.373 2 gap/(bw + bH) = 0.0995ta = 0.325 BH = 3.81, eH = 0.85, bwbH/Sw = 2.236 32 1 tal tail loads corresponding o the various tailplane in-

c =3(, - c a = constant

LH 'H~H " t a i l off mtail on A comparison o f the computed tail on drag for the complete airplane with the experimental drag shows that the cross induced drag term is very important when the tail is carrying a download; i t s calculated value over- comes the skin friction drag of the tail assembly and the self induced drag o f the at C L ~ = -0.4. The experimental data do not quite confirm horizontal tail itself this result: The experimental skin friction of the tail assembly and the self in- duced drag of the horizontal tail were underestimated; but note that the general shape of the complete airplane drag with tail plane lift curve i s correctly predicted.

It is concluded that biplane theory presents a simple method for calculating tail drag, and that the trim drag penalties are generally small, for foreplanes or tailplanes of reasonable span and loading.

A -

1.

!

I

i

I i .

4 : i 1 .

I i ' v . .

, .

lr' 11 ' I L Y x 0 .

I 0 )

e

n

-

.-

I -

-

E C Q) E

.-

L Q)

ft

w N

0 '

3 I

8. ss

8.1 Overview of Configuration

Part I - Cost Considerations for Aircraft Configuration Changes

R. Tuhl inson, Beech Aircraft Corporation

Part I1 - Aerodynamic Considerations

J. Roskam, University of Kansas Learjet Model 25 Drag Analysis 8.2 R . Ross and R. D. Neal Gates Learjet Corporation Problems in Propulsion System Integration 8 . 3 Research Center

W. Henderson and J . Runckel, NASA Langley

8.4 Determination of the Level Flight Performance of Propeller-Driven Aircraft

E. J . Cross, Jr. Mississippi State University

R, R , Tumlinsan Beech Aircraft Corporation Drag reduction, i n our industry, i s a principle that ranks with Motherhood. There are about as many aircraft engineers who would demean ways to reduce drag as there probably are politicians who would attack apple pie. Most of the people here have spent a great deal of time searching for ways of reducing drag and of trying to convince others of the merits of the efforts required to do so. I am sure that I have a lot of company i n the frustration that goes with that search and effort.

We are al I members, or supporters, of an industry which i s fueled by profit. And that profit is directly dependent upon delivery of aircraft which provide good value for our customers. Or, simply said, the costs of changes and evolutionary improvements must be at least balanced by the benefits.

I'm not an expert by any means on costs, but I have had a lot of experience in.

trying to overcome the obstacles provided by the cost considerations of proposed changes.

S o , today, let me take the role of the Devil's Advocate on aircraft costs and cite some of the considerations which most be made and which may outweigh potential performance improvements.

As a means of illustrating both costs and benefits, I'd like to present a very arbitrary example which wil I touch on many of the important cost considerations which must be made to arrive at a production decision.

Let's say that we have arrived at a modification which will reduce the drag of our airplane so as to provide an increase in cruising speed of 4 mph. In the course of this workshop, we have considered many possibilities for achieving this, so I'm not going to specify how this was achieved. But, as an arbitrary assumption, let's say that we can, in fact, increase our cruising speed by 4 mph; and also, just as arbitrarily let us assume that the resulting changes would net an increase in airframe weight of 5 pounds. This represents a loss of payload of 5 pounds. And, in addition, this will typically require design and production changes to another 10 pounds of the existing airframe weight, although this would depend in a particular case on the nature of the configuration change.

For the purpose of this example, we will apply this to a hypothetical turboprop with a CN ising speed of 250 mph .

Precedin Page Blank

33 1 am using have been chosen to suitable for this ex First, let's consider the cost increase for direct labor.

he 5 pounds of new and increased weight will be a net increase to the airplane, and an appropriate slope and man-hour/pound figure must be selected for the new effort. The 80% learning curve i s found to be fairly typical for the general aircraft industry,and I believe t for this example. The 5 pounds will be projected from Unit 1000 at 1 .O to obtain Unit No, 1. This is shown to be an increase of 46 hours at Unit 1 with a cum- ulative increase of 15 hours for 100 units.

A somewhat different consideration must be made for the "changed" weight of 10 pounds, where there i s not the same potential for "learning" improvements. Something less than the 80% slope typical for ''new" production would be more appropriate. If we assume a 90% learning-curve slope, using the same factors as before (1 .O man-hours/ pound at Unit 1000 for 10 pounds, this time), the cumulative man-hours over 100 units is 1 6.6 man-hours.

As the changed effort has replaced an existing task at 1 .O man-hours/pound, this 10 hours can be subtracted from the 16.6, leaving 6.6 hours for new learning. Our

total dire'ct labor increase now becomes 15 hours + 6.6 hours, o r a total of 21.6 hours

each for the 100 units.

Next, these man-hours must be converted to dollar costs. The latest figures pub1 ished by the Bureau of Labor Statistics show that direct labor rate applicable for the aircraft industry as a whole i s $5.78 per man-hour. With inflation and differences within the industry, this rate can become obsolete quickly. Overhead and direct expenses plus general and administrative expenses can add a so-called "burden" of 200 to 300 percent to this rate.

TABLE I Overhead and kect Expenses Include Indirect salaries and wages; support labor such as planning, scheduling personnel, etc.; holidays, vacations, sick leave, etc.

Insurance, payrol I taxes , social security, group I ife-insurance, workmen compen-

sation, retirement plan, sales taxes, personal property taxes, and depreciation.

Maintenance and repair on shop equipment and on buildings.

Shop supplies such as perishable toots, office supplies, etc.

Travel, telephone, freight, etc.

Overtime premiums, product liability, etc.

General and Administrative Expenses Inalude Executive and management scilaries; accounting; procurement; office suppi ies;

and other costs which cannot be directly associated with labor cost - either

manufacturing or engineering.

There are two other contributions to the costs: the materials and the develop- ment costs. Material costs for an airplane in this category vary with the size and com- plexity. Development costs also vary with the class of airplane and the accompanying complexity of the development effort and the FAA certification program required. A range of $3000 to $4000 per pound i s the general ballpark figure when everything is added up, and in this example 100 units was selected to amortize these costs.

When the pieces are all assembled, a price change can be determined for the improved airplane which adds up to approximately $1 600.

The cost to the customer must be weighed against the additional value to the customer. Let's look at it this way: our hypothetical airplane cruises at 250 mph. To keep it simple, I'll use this cruising speed, although it would be more accurate to deter- mine an average block speed based on a customer's particular routes. For a customer's usage, we will assume 600 hours per year. Appropriate operating costs are quoted on Table 1 1 . And, as noted on the table, the costs of the modification can be recovered by the savings i n operating costs i n 1.87 years. After that time, the savings would represent a net gain to the customer which would continue.

perating Cost Co~par~son efore Modification Direct Operating Costs/Hour $ 77.50 13.18 Indirect Operating Costs/Hour $ 90.68 Total Operating Costs/Hour Cruising Speed 250 MPH Total Operating Costs/Mile = $90.68/HR $ 00.3627 250 MPH After Modification $90.68/HR Total Operating Costs/Mile = $ 00.3570 254 MPH Savings Per Mile 0.0057 Savings Per Year = $.0057 x 250 MPH x 600 Hours = $055.00 $1 600.00 = 1.87 years $ 855/YR I am not going to exercise judgment on the 1.87 years, because of some of the arbitrary assumptions that could drastically affect the results. In estimating costs for a particular project, appropriate values must be used which would not be the same as those used in the examples. The actual special improvement, the weights affected, the cost factors that are current for a particular project, the number of units used to amortize the development costs could each produce significant differences. I believe the figures here are representative, but primarily, I hope they illustrate the key factors which can affect a production decision considering the costs involved.

And, even after this type o f analysis, there are other factors which may strongly influence both the costs and/or the decision to proceed. For example, the FAA recerti- fication considerations. If this can be avoided, perhaps by timing the modification with a complete model change, these cost figures would look much more favorable.

In the last analysis, a decision to proceed may depend on the philosophy of the company management. Beauty i s in the eye of the beholder, the saying goes. There are many changes made in the name of progress, or to satisfy a dedication for a clean-looking be difficult to justify solely on the basis of this type of comparison.

Id not be desira le. But--- sometimes the stroke of

ugh to convince a cost-minded management --- and that*s where this

analysis would help.

Thank you.

of the ~ o m p ~ e t e Configuration Jan Roskam University of Kansas Introduction The purpose of t h i s part of the paper is to focus on a numb relate them to the performance of the complete configuration.

First, the effect of fuselage camber, wing and nacelle incidence are discussed from a viewpoint of design decision making.

Second, the effect of overall cruise drag on the design gross and empty weight of the airplane i s discussed. Examples show that cruise drag can have a very important influence on total airplane weight.

Third, the effects of usable cruise lift-to-drag ratio and wing-loading are shown to be important.

Finally several research needs relating to design of the complete configuration are reviewed.

Effect of Fuseloae Camber, Winn and Nacelle Incidence In putting together a new airplane, a number of fundamental geometric choices must be made. Typical exampies of such choices are: - extent of fuselage camber;.

- wing incidence on fuselage; and

- nacelle incidence and p o s i t i o n relative to the wing.

In determining the extent of wind tunnel testing required to "optimize" the configuration, the aerodynamicist is confronted with a large number of variables. For example, if it i s assumed, that two camber shapes, two wing-fuselage incidence angles and two wing-nacelle incidence angles are to be investigated, this alone leads to eight combinations to be lested. Under the economic constraints of the general aviation industry it is usually not feasible to d o this much testing.

Major aircraft manufacturers, on fighter, bomber and even on some trans- port programs, obtain significant inputs from NASA in terms of systematic wind tunnel configuration testing.

How does the general aviation designer choose the best configuration? Well, very often he ends up guessingor, the shaping decision (for lack o f definitive .

aerodynamic input) is made for him by engineers or managers outside of aerodynamics.

Preceding Page Blank

weigh the aerodynam~cist ling costs and market~~g opini cause the aerodynamic~st does not in the decision ~ a ~ ~ n g process, have convincing argumen~s one way or the other.

To illustrate these points and to point once more to the need for systematic tunnel testing of general aviation research models, the following examples are given .

Note from Figure 1 that three different vertical nacelle install used for turbopropeller airplanes. Note also, that a l l three use rather differing aft fairing shapes. The question arises: can they a l l be right?

Observe from Figure 2 that one manufacturer employs two quite different piston engine nacelle configurations. Figure 3 illustrates two more and again different nacelle shapes. The questions arises again: can they all be right?

Possible pay-offs of such research are illustrated in Figure 4 taken from Reference 1 (1942). Figure 4 shows a range of wing-body-nacelle drag coefficients of .1250 to .1050, (.0078 to .0066 based on wing areal) depending on vertical nacelle location alone. In other words, there are 12 drag counts to be gained by selecting the vertical nacelle location.

It would seem that the industry could derive significant benefits from a series of systematic wind tunnel test to determine the best (lowest drag) shape o f such wing-nacelle installations. Such research should also account for the effect of ion and orientation, as well a s for the possible beneficial effect of ' forward propeller shaft extensions, such as used on the Navajo.

Drag Effect on Airplane Weight and Airplane Market Price Aerodynamic drag i s not generally thought of in generaf aviation airplane design as an important factor affecting airplane weight. The reason may be the fact that usually new airplane "designs" consist of adaptations of components which are already in production, to a new airplane. The term "tinker toying", although not a kind description probably applies to much of general aviation airplane design.

However, every now and then a truly new design evolves and then the effect of drag on weight can be important as w i l l be illustrated in the following simplified analysis.

Assume that total airplane weight i s broken down as follows:

wF + w

w = WPL f

E where: = pay~oad weig = fuel weight ( ~ n c l u ~ i n g reserves) W, WE = empty weight Fuel weight and empty weight are assumed to be broken down as follows:

-

WF = A f TxSFC x - and: (2)

R

-

WE = B +zT + E W F

(3) where:

-

A = weight of reserve fuel T = cruise thrust SFC = cruise fuel consumption Ibs/lbs/hr V cruise speed R = cruise range

-

B = empty weight without power plant and fuel system

C = weight of power plant per Ibs of cruise thrust

-

D = weight of fuel system per lbs of cruise fuel In cruise flight: T = W a n d L = D (4) lift drag so that D T = W ( r ) (5) Substituting equations (2) through (5) into equation (1, yie.Js:

w = ~ ~ , , ( ~ + ~ ~ ( S F C ) ~ ~ + C 8 + Z W ~ ) +

c6)

k

it is possible to rewrite equation (7) as: CL

VJ = -b/L/p

To determine the effect of drag on airplane weight, the differential ' s w I ; ) ~ ~ p ) can be found from equation (10) as: Table 1 presents data from which can be calculated for a typical general aviation piston engine driven twin.

So using equation (1 I): This means that per unit L/O, the airplane gross weight can be lowered by about 120 Ibs, a significant saving when compared to ihe empty weight, WE.

Figures 5 and 6 illustrate similar results obtained in Reference 2 on small two-place turbofan (1 200 Ibs max thrust) airplanes.

Table 'I and Figure 5 and 6 all show the importance of designing to the maximum possible cruise lift-to-drag ratio, if the lowest possible airplane weight is to be achieved, It should be noted, that lower empty weight, achieved by better aerodynamic design has a very significant effect on the marketing price of an airplane. Table 2 shows typical market prices related to gross and empty weights for general aviation twins.

For the example twin of Table 1 the typical market price per pound of empty weight would be about 34 $/I&. Attaining a 120 Ibs saving would cut the market price by $4,080, a rather significant competitive advantage!

= 3700 l b s Engines 2 x 300 hp, a t 450 l b s each E WF = 1000 l b s WpL = 1600 l b s SFChp = .45 l b s / h p / h r s W = 6300 l b s Assume propeller and engine weight =1100 l b s Assume fuel system weight = 100 lbs

Assuming a cruise L / D = 11 and Wave - - 5,800 l b s

cruise Tcruise = 527 lbs = 216 mph, t h e n HPcruise = 303 "cruise Fuel flow i n c r u i s e then is 136 lbs/hr. T h i s yields a range o f 1ooo-200 (reserves)) 216 = 1270 miles. The value o f SFC i s (

13' - .26 l b s / l b s / h r

327- -

S o , A = 200 lbs B = 3700-1200 = 2500 l b s

- 1100 E = 100

= -13 c = = 7 J - = 2.1

1,000-200 From equations (8) and (9): a = 1600 + 200 (1 + .13) + 2500 = 4326 lbs 21 6 b = -26 x 1 ~ 6 (1 + .13) + 2.1 = .05 + 2.1 = 2.15 Lift-to-Drag Ratio and Wing Loading Effects Revisited as the Cessna 172 typically cruise at lift coefficients Light airplanes, such in the range of:

C L < -3 -to - s

Figure 7 shows that the corresponding L/D value varies from 10.0 to 13.2. This compares with a maximum L/D value of 13.8 indicating that significant improvements must be attainable by increasing wing loading. Increasing wing loading not only w i l l bring the cruise CL closer to L/D/max on the polar but i t w i l l also shift the polar to General Aviation tight Twin irframe Prices, - Price Price

Gross Weight I Empty Weight

.-R- $ Type W (lbs) W E (1bs) $/1 bs -- $/Ips 1 Cessna Skymas ter 4,630 2,710 63,300 13.7 23.4 !

23.1 i

Piper Seneca 4,570 2,770 63,995 14.0 , Piper A t t e c E 5,200 3,042 88,200 1 7 . 0 29.0 Beech Baron 855 5,100 1 7 . 5 28.2 3,155 89,000 Cessna 310 5,500 3,251 89,950 76.4 27.7 --.

5,000 I 2,986 78,889 , 15.7

Averages Rockwell Shri ke Comnander 4,608 6,750 7 28,150 1 9 . 0

I

Cessna 402 B 6,300 3,741 138,500 2 2 . 0 Piper Navajo B 6,500 3,930 139,100 21.4 Cessna 414 6,350 4,042 174,950 2 7 . 6 43.3 6,775 4,265 219,450 32.4 Beech Duke

6,535 ! 4,117 160,030 24.5

Averages range (note that C actually w i DO areal) This fact has been previo demonstrated also in such papers a 5 References 3, 4, and 5.

Figure 8 illustrates some typical results. Cutting wing area in the chordwise direction by 30 percent results in a 10 percent reduction in thrust required and therefore in fuel flow. Figure 9, shows the relative aerodynamic "cleanness" of 1975 general aviation single engine airplanes compared to what i s feif feasible in the future. To achieve this however, w i l l require the introduction of new designs and new manufacturing technology.

Research Needs It appears that research into the following areas would have significant potential for paying off in imporved general aviation airplanes:

- Nacelle shape and nacelle location on wings (for horizontally opposed

piston engines and for turbo propeller instal lation);

- Improved methods for predicting the effect of drag on weight (Adaptation

of NASA/Ames GASP?); and

- Expansion and specialitlation of GASP to single engine and twin engine

propeller driven airplanes with detailed accounting for weight, stability and control and propulsion interference factors.

References 1 . Becker, G.V. and Leonard, L.H.; High'SpeedTests of a Model Twin-Engine Low-Wing Transport Airplane; NACA .TR 750, 1942.

Heldenbrand, R.W.; Merrill, G.L. and Burnett, G.A.; Study of Small Civil 2.

Turbofan Engines Applicable to Military Trainer Airplanes: Final Report; NASA CR 137575; Garrett Airesearch 74-210987-A; April 1975.

Roskam, J.; Opportunities for Progress in General Aviation Technology; AIAA 3.

Paper 75-292, Presented at the AJAA 1 lth Annual Meeting and Technical Display, Washington, D .C. February 24-26, 1975.

Roskam, J.; New Airfoils and Higher Wing Loadings: A New Look at General 4.

Aviation Airplane Deqign; Paper presented at the Technological University of Delft, The Netherlands, May 20, 1974.

Roskarn, J. and Kohlman, D.L.; An Assessment of Perforrnancy, Stability and 5 .

Control Im rovernents for General Aviation Aircraft; SAE Paper 700240 pre- sented at t E e SAE Business Aircraft Meeting, Wichita, Kansas, April 1970.

-.

Beechcraft SS Airliner r~ven!ecn-rea1 light transpod

I A

n

I

P i p r PA-31T Chryilnnr rirjright-seat light transport aircrd (Pibr Pnu) Figure 1 Examples of General Aviation Turbopropeller Installations I T i m - v i e w drawing of the Beechcraft Baron 58 four/six-seat cabin monoplane b r r r - r l e w drawlng 01 the Beechcraft Duke B60 4p-seat prrrrofird transport (two 380 hp Lycoming TIO-541-ElCQ engines) Figure 2. Different Piston Engine Nacelle Shapes Used by O n e Manufacturer

I

1 i

Cessna Model 414 prossurised light transport L U Three-view drawing of the Piper PA-31P Pressurised Navajo (two 425 hp Lycoming Ti60-541-E1A engines) Figure 3. Further Examples of Piston Engine Nacelle Shapes Figure 4. Example of the Effect of Vertical Nacelle Location on Drag c x c

.-

>

.-

c

.-

E 0 ) c)t 0)

E

D + % f ,o

I3

+ x +

.-

> .C c

.-

u) c Q v) C

.-

.a-

P

c W a , In

.-

G

aJ N

.-

v) a J t

.-

0 ) c W Figure 7. Typical Single Engine Airplane Drag Polar I . 2 . 4 .6 . .0. 1.0 CHORDWISE F R A S T I O N - O F .

WIN6 AREA x Figure 8. Effect of Chordwise Area Reductions on Thrust Required 2000 3009 600 WQO

TOTAL WETTED A R E A - FT2

Figure 9. Typical Range of Wetted Areas and Equivalent Parasite Drag for General Aviation Airplanes and Ronald D. Neal Gates Learje t Corporation Drag Analysis The procedures and data for estimating drag at Gates Learjet are contained in the Learjet Aerodynamics Handbook and were used to calculate the drag charac- teristics of the Model 25 airplane. Based on cruise flight test data obtained on the Model 25, these methods generally predicted the total drag characteristics within current acceptable and reasonable engineering accuracy The use of wind tunnel model results will not guarantee absolute accuracy because of the many corrections and interpretations that must be applied to the data.

However, small scale tunnel tests can provide the technique for minimizing con- figuration drag as well as identifying the aerodynamic contributions of each individual component.

Flight testing, when carefully executed, will provide the complete trimmed drag of the airplane. Such a program requires extensive testing since it i s necessary to define the characteristics throughout the operating envelope of the airplane.

What a flight program does not do and cannot do (within practical limitations) i s to isolate and identify drag characteristics for each of the major components of the total vehicle. Without knowing the drag build-up for the airplane it i s difficult and costly, from flight t e s t data alone, to identify drag problems and then through the continued use of flight tests to arrive at a solution to the problem.

Only by integration of the results of a l l the available techniques can con- fidence i n drag prediction and eventuaf control o f drag levels be developed.

The total airplane drag i s produced by several separate contributions that are identified as: profile drag (skin friction) interference drag

- roughness and gap drag

* induced drag compressi bi I i ty drag

Preceding Page

* profile drag variation with lift trim drag s then reasoned that from these sources an deter~ined e The f o ~ l o w ~ ~ g com~ents provide the reasoning and analysis for using these data to determine the drag for the Model 25. For purposes of evaluation and comparison a mid-cruise weight of 12,000 pounds, an altitude of 40,000 feet and a cruise Mach number of 0.75 w i l l be used.

Figure 1 presents the trimmed, level flight drag characteristics for the Model 25 at cruise Mach numbers of 0 . 6 0 , 0.70, 0.75 and 0 . 8 0 . For a weight of 12,000 pounds and a cruise Mach number of 0.75 the lift coefficient i s 0.336 and the corresponding total drag coefficient i s 0.0338.

Profile, Interference, and Roughness Drag The estimated skin friction drag i s 0.0186, interference drag i s 0.0032, and roughness and gap drag i s 0.0016. The total zero lift drag i s then estimated to be 0.0234 or 69.23 percent of the flight test cruise drag. Therefore, i f the zero lift drag estimate i s correct, the balance of the drag, 0.01011. may be attributed to: induced drag compressibility drag 0 profile drag variation with lift trim drag Induced Drag One accounting technique that can be used in evaluating the drag contribu- tion due to induced drag i s to evaluate the induced drag term with the span efficiency factor equal to 1 . O . By using this procedure all of the losses due to non-elliptical spanwise loading and wing-tip tank effects w i l l be included in the profife drag Using this technique the induced drag at the cruise condition variation with lift.

i s 0.0072 or 21.30 percent of the total cruise drag. For reference purposes, Figwre 2 presents a plot of induced drag based on e = 1 .O.

At this point the contribution due to zero lift drag and induced drag is 0.0306. The remainder of the cruise drag 0.0032 or 9.47 percent should be accounted for by 0 compressibility drag profile drag variation with lift 0 trim drag .02 .03 .04- Q5 CD Figure 1 . Learjet Model 25 Drag 0.7 Q.6 0.5 0.4 .0.3 O.L 0.1 Figure 2 . Learjet Model 25 Induced Drag ased on f~jght test data, Figure 3 presents the comp ibility drag incre- sing these data the compressibility drag coe Mach number of 0,75 and a lift coefficient of 0,336, i s determined to be 0.0028.

The r e ~ i n i n g drag increment of 0.0004 should be the sum of the profile drag variation with lift trim drag I n comparing the actual flight test compressibility drag increments to the original estimated curves it was noted that the flight values were higher than the predicted values. The difference between the actual and the estimated increases with Mach number and lift coefficient which i s not unexpected. The reason for this ' difference may be better rationalized i f the prediction procedures are reviewed.

As previously noted the total drag of the airplane may be attributed to profile drag (skin friction), profile drag variation with lift, interference drag, roughness and gap drag, induced drag, compressibility drag and trim drag. At zero lift the induced drag contribution i s zero and the remainder of the zero lift drag should be accounted for by the other contributions.

Considering the data presented in Figure 4 and similar data for M = 0 , 6 , the increment between the zero lift drag coefficient values for M = 0 . 6 and M = 0.75 should then be equal to the compressibility drag and the trim drag contributions. The difference between trim drag at zero lift for these two Mach numbers wi I I be considered as being insignificant rn The reason for 'this assumption i s that between these speeds the stability values that determine trim drag should not be significantly different. Therefore, the increment between the zero lift drag coefficients should represent compressibility drag alone.

At zero lift for M = 0 . 6 , CD = 0.0210 and for M = 0.75, CD = 0.0226 with ACD = 0.0016. From the data of Figure 3, the compressibility drag increment at CL = 0.2 i s determined to be 0 . 0 0 1 5 which i s in good agreement with the sixteen count increment at zero lift. This correlation would then indicate that from M = 0.6 to 0.75 the compressibility drag increment i s the same for all lift coefficients in the range from 0.0 to 0.2. Swch results are not unexpected with similar trends being shown in available literature. At Mach numbers greater than 0.75 the compressibility drag increments for lift coefficients of Om0 and 0.2 should deviate as shown.

. 6 0 -6 5 .70 .7 5 -80 -85 MACH N U M B E R Figure 3. Learjet Model 25 Compressibility Drag - M = 0.15

' O ~ F L I G W T Te17 DATA /

I 1 1 a20 -02s a30 .055 -040 9045 a50 =n Figure 4. Learjet Model 25 Drag tunnel test resu t s and letting e = 1 . O for e drag ~ a ~ ~ a t ~ o n with lift i s d e t e ~ m i n e ~ from C D ~ = CD - Cgi with the results being presented in Figure 5.

Using the data of Figure 5 the profile drag increment due to lift for a CL = 0 . 3 3 6 i s ACop = 0 . 0 0 0 7 o r 2 . 0 7 percent of the totaf cr se drag. However, this value i s 0.0003 m r e than the total drag increment a l l for both profile drag variation with lift and trjm drag.

Trim Drag In considering the trim drag increment it is noted that the basic skin friction drag of the horizontal tail has already been included in the basic profile drag of the airplane. A profile drag variation with lift will exist for the horizontal stabi- lizer.

However, this contribution is probably s o small as to be negligible. Thus, the horizontal stabilizer t r i m drag increment will be considered to consist only of the induced drag contribution of the tail.

Low speed wind tunnel test data are used to show that the tail induced drag or trim drag increment i s 0.0005. Compared to the total cruise drag of 0.0338 the trim drag amounts to 1.48 percent.

I t i s noted the sum of the estimated profile drag due to lift, 0 . 0 0 0 7 , and the trim drag, 0.0005, i s 0.0008 more than the total drag increment allowed for them.

of Drag Analysis Discussion By using the reasoning and procedures given i n the preceding analysis, a l l of the drag, except for a drag increment of 0.0008 can be accounted for in the analysis. Balance of the total drag picture can be obtained by slight revisions in the estimates of any of the individual sources. However, the more likely and suspect areas for reassessment are the contributions due to profile drag (skin friction), interference drag, and roughness and gap drag. These items are open to question because they represent the estimated portion of the previous analysis, whereas all of the other items have a firmer basis for conviction. The eight drag count increment represents 2.37 percent of the cruise drag. In order for the individual drag contributions to balance it i s reasoned that this drag reduction may be distributed between profile, interference and roughness so that the total for these three sources i s 0.226 instead of the original estimate of 0.0234.

f the zero lift drag is 0.0226 it s ible to take the flight = 0,75 ( ~ ~ g u r e )c plot these da o lift and verify the 0,0226 value, of C L ~ versus CD with the symbofed points being taken directly from Figure 1.

The zero lift drag, as determined from this plot is 0.0226 which then substantiates this zero lift drag value as determined from the previous analysis, Distributing the eight drag count reduction between the three sources s o that the total zero lift i s 0.0226, the breakdown of total airplane cruise drag i s then given in Figure 6.

Based on a total profile drag of 0.0180 and on the original estimated pro- file drag contribution of each individual component the profile drag (skin friction) may be summarized as shown in Figure 7.

The total profile drag accounts for 53.25 percent of the total cruise drag of the airplane.

Drag Distribution Figure 8 presents the drag distribution for the airplane as a function of Mach number with the data being extended.to Mach 0.85. This plot provides a summary presentation of the drag contributions of the various drag sources discussed i n this report. Throughout the flight range the drag contribution due to profile drag continues to be the major source of drag representing 61 to 66 percent of the total cruise drag. With increasing cruise speed the induced drag decreases, varying from 24 to 11 percent of the total. The compressibility drag increases with increasing Mach number, varying from 4 to 24 percent. The contribution due to trim drag and profile variation with lift represents the smallest source for a range of 8 to 2 per- cent of the total cruise drag.

A comparison of high Mach number estimated drag with flight test deter- mined drag i s presented in Figure 9.

0.t Figure 5. Learjet Mddel 25 Profile Drag Variation with Lift M = 0.75 CL = 0.336 CD = 0.0338 Source X of Total

-

Profile drag (skin friction) .0180 5 3 . 2 5 Profile drag variation with lift

. MI07 2.07

Interference drag -0031 9.17 Roughness and gap drag .0015 4.44 Induced drag A072 21.30 Compressibility drag .0028 8.28 Trim drag .W05 1.48

- -

TOTAL 0.0338 100.00 Figure 6. Cruise Drag Breakdown CQ,,, Profile Drag = 0.0180

Item r o f ACQp

-

ACDp

-

Wing 29.57 ,0053 Fuselage 34.95 -0063 Tip Tanks 11.83 .w21 Tip Tank Fins 0.54 .mol Nacel 1 es 6.45 .0012 Pylons 1.61 .0003 Horizontal 4.14 .0016 Vertical 5.91 .0011

-

TOTAL 100.00% 0.0180 Figure 7. Profile Drag (Skin Friction) Breakdown 36 1 1 0 1 f 0.70 0.75 0.80 0.85 MACH N U M B E R Figure 8. Learjet Model 25 Drag Distribution EXAMPLE ( 3 MID-CRUISE WT.

40,000 FT.

% M CL CD CD ACD D I F F .

(Actual ) (Estimated) .70 .385 0353 ,0353 0 0 .75 ' 336 .0338 .0335 .OOO3 0.9

.77 .318 - 0330 .0331 .OOO7 2.1

.80 .295 .0341 .0326, .0015 4.6 Figure 9. Comparison of Flight Test Drag and Estimated Drag

enderson a n d J . Runckel

Langley Research Center Problems in Propulsion System Intenration The problems associated with propulsion system integral Ion a r e related to the placement of the engine on the airframe. The complexity o f the problem, and the associated drag, depends o n the area where the power plant is located--as indicated on Figure 1.

Wing installations c a n consist of nacelles located under or over the wing or propulsive wing concepts. The exhaust stream can produce either favorable or unfavorable induced effects o n lift and drag.

The fuselage could contain lift engines or a d j a c e n t nacelles which produce interference drag a n d t h e hot exhaust c a n have detrimental effects o n t h e structure.

The afterbody with buried engines i s a particular problem with military aircraft since a large portion of the drag occurs in this region.

Reverse thrust o n rear mounted engines can pose plume impingement problems o n stabilizing surfaces.

I wilt touoh briefly on several of t h e areas indicated on Figure 1 .

The main problams with jet-engined aircraft a r e with the induction system and the exhaust system and t h e effects on configuration performance consisting of mutual interferences between the propulsion system and the airframe.

Interference effects a r e usually evaluated first by looking a t the isolated’ components of the propulsion system and then with the system integrated into the airframe. I would like t o present some recent NASA work o n jet engine components which c a n influence t h e related aircraft drag areas.

Preceding Page Blank

F O REBODY

PROPELLER/ENG I NE, I N E T S

NING

FLOW Fr E L D / ~ C E L L E I N T E R F E R E N C E

I N D U C E D EFFECT - L I F T AND DRAG

FUSELAGE

INTERFERENCE DRAG .

EXHAUST EFFECTS

INDUCED PRESSURES - S T A T I G / D Y N A M I C - ~ C O U S r I C

I N D U C E D TEMPERATURES

AFTERBODY'

POWER EFFECTS ON DRAG.

Nozzt~/A I RCXAFT CLOSURE

REVERSE THRUST

EMP E id N A G E

PLUME EFFECTS O N STABILITY

VERTICAL TAIL ENGINE INSTALLATICN

Influence of Propulsion System on Airframe Figure 1 .

Aerodynamics inlets has recently been extended to higher u ~ b e r s and updated for geometries corresponding to the mass flow require- ments of high bypass r o b r fan engines. Figure 2 shows the experimental results o f the investigation. The variables were fineness ratio, inlet highlight to maximum diameter ratio mass flow and angle of attack. The upper left chart i n Figure 2 shows the effect of fineness ratio a t a constant mass flow ratio of 0.8. The short inlets are best at lower speed but drag rise Mach number increases with fineness ratio. Figure 2 shows the effect of diameter ratio-drag increases as the inlet i s opened up but the critical Mach number i s extended. The right side of the figure shows the effect of mass flow ratio variation where reductions i n drag and increases i n the knee of the drag curve increase with increased inlet mass flow ratio. These data were obtained from force balance measurements and the drag coefficient i s based on the maximum cross-sectional area.

Reference An Investigation of Sever NACA 1-Series hisymmetric Inlets a t Mach Numbers from 0.4 to 1.29. Richard, J. Re. NASA TM X-2917, March 1974.

W P cu 9 9 9 Q N

9 4

Q ressure destribution measurements were also obtained during the inlet ~ n v ~ t i g a ~ ~ o n . ~ ~ g w r e 3 shows a comparison of a theoretical prediction with on experimental pressure distribution. The measured pressure coefficients are indi- cated by the circular symbols for both the outer cowl and on the inside of the inlet.

Data are for an inlet with highlight to maximum diameter ratio of 0.85 and fineness ratio of 1 .O at 0.7 Mach number and a mass-flow ratio of 0.87. The calcutated pressure distribution using a stream tube curvaiure theory developed for NASA i s shown by he line. This theoryaccounts for both the internal and external flow field at the inlet and can also be used to calculate the pressure distribution on afterbodies with a jet exhaust. The agreement i s very good with only Q slight miss on the I ip peak suction pressure.

Reference Analystical Method for Predicting the Pressure 0 istribution about a Nacelle a t Transonic Speeds. J . S. Keith, D. R. Ferguson, C. L. Merkle, P. H.

Heck and D. J. Lahti. NASA CR-2217, July 1973.

. .

. .

.,. . ; .

. I m e i l d an extensive data base on the drag characteristics of A series of nozzle boattail configurations has been investigated at sub- sonic and transonic speeds to determine the pressure drag for various shapes, fineness ratios and closure ratios which should be applicable to several type5 of jet and fon engines. Shown i n Figure 4 are samples of data for circular arc boattails from a program where fineness ratio varied from 0.8 to 2.0 and closure ratio from 0.5 to 0.7. The variation o f drag with Mach number i s shown for a short steep boattail on which the flow separated and a higher fineness ratio boattail with attached flow. Experimental data i s indicated by the symbols and theoretical predictions by the lines. For attached flow the prediction i s reasonable until supercritical velocities are approached. For the separated flow case, how- ever, existing theory i s inadequate and further work is being done to improve the situation. The limitations to existing theory include transonic flow and imbedded shocks, separated flow regions and inadequate modeling of the jet exhaust to include entrainment effects.

37 1 u)

d

II

.-

c a cr:

n 9

- i!

C I U 0-

.-

I" c

\"

(r

E

.-

" P

L

L .-

I

E c tl

i

m

U I

U

I I . I I

d c \ I QD d- 0

-

L

0, P

En

.-

LL ing 1nsta.aliation Figure 5 shows an example o tern completely integrated into the wing. The configuration i s a four tip-turbine driven fans. The wing fans, oriented tion, induce air into the leading edge of the wing and exhaust i t at the trailing edge. Either l i f t and/or axial thrust are obtained by means of a sli deflection and modulation In the cruise mode, varia as well as provide e flow from the gas gen the wing was geometrically thick (thickness to chord ra aerodynamically it was relatively thin at a Mach number o f 0.85.

Reference Longitudinal Aerodynamic and Propulsion Characteristics of a Propulsive- Wing V/STOL Model at High Subsonic Speeds. Leland B. Salters, Jr., and James W. Schmeer. NASA TMX-2693. January 1973.

~pper-surface blow in^ propulsive l i f t concepts have shown a potentia4

eed high-lift performance with lower noise I cause of the shielding afforded by the wings between the jet related noise and the ground. Only preliminary work has been done on the effects of propulsive lift cruise performance. Shown on Figure 6 are results from an experimental and analytical investigation to determine the effects of forward mounted jets blowing over a wing. The model had a low aspect ratio wing and forces were measured on the wing-afterbody portion. The nacelles were located forward and above the wing. The analytical method represented the wing lifting surface with a lattice of horseshow vortices and simulated the effects of the exhaust plume with a line sink-source distribution located on the axis of the jet. The theory accurately lift due to the jet flow for Mach numbers predicted the variation i n interference of 0.4 to 0.7. The theory also predicted the reduction i n induced drag indicated by the data. The favorable increment in induced drag increases with both the

ratio of jet velocity to free stream veiocity and - wing l i f t coefficient. The plots

on the right side of Figure 6 show that the theory correctly predicts both the increase i n lift and the reduction in induced drag coefficient.

References Exploratory Investigation a t Mach Numbers from 0.40 to 0.95 of the 1 .

Efects of Jets Blown Over a Wing. Lawrence E . Putnam. NASA TND- 9369, November 1973.

An Anaiytical Study of the Effects of Jets Located More than One Diameter 2.

Above a W i n g a t Subsonic Speeds. Lawrence E. Putnam. NASA TND-7754, August 1 974.

CT .-

Q)

a-

a U M d o ( I d - 0 "

u

a

we developed was also used t o correlate some e Fokker VFW research results presented a t a 1974 AGARD meeting are shown in Figure 7. The engine-nacelle configuration i s shown i n the upper right hand corner and was located above and forward of the wing. The plot shows the change in induced drag coefficient for jet off and a jet to free stream velocity ratio of about 7.5 by the data points; circles for cruise, squares for climb values of left coefficeint. The theory calculationsare indicated by the broken lines, and again, the predictions agree well with the data. The actual values of velocity ratio required for climb and cruise are indicated by the broken bars. The reductions i n induced drag at climb are’stifl substantial and some benefits s t i l l occur at cruise.

Reference Airframe-Engine Interaction for Engine Configurations Mounted Above the Wing. Part 1: Interference Between Wing and Intake/Jet, by G.

Kreng . In: AGARD Conference Proceedings No/SD on Airframe/Pro-

pulsion Interference AGARD-CP-150, September 1974.

.

r9

P d

I I

/

w

' J

I

/ o

_.

/

a

I

/

/

ru

/ / / t- I m.

z

W lllllilil

E

I

EOCf

w

n

X

w

111111111

I

).

Z1 W I t -

.-

I on Cririse Efficiency hese exploratory results have encouraged us to look at the performance b e n e f ~ ~ that may be sible with transport configurations during cruise flight , igure 8 illustrates a current program aimed at determining the cruise efficiency of highly integrated propulsion systems. A powered model of a basic transport configuration i s under construction which w i l l allow various types of engine locations on the wing to be studied and performance comparisons to be determined. Three types of engine installations will be considered: conventional under the wing pylon mounted nacelles; over the wing nacelles; and a blended upper surface blowing configuration. An internal balance system wil I measure all forces on the model including the thrust. The types of propulsion systems are shown on the right and consist of flow through nacelles, blown nacelles with air jets and turbofan simulators. These simulation systems wil I provide aerodynamics modef data comparisons, the jet induced l i f t and drag increments and the coniri- bution of the inlet flow simulation on the flow field.

tion presented on Figure 9 shows the drag i ~ c r e ~ e n ~ that can be expected with an aft fuselage mounted nacelle configuration, represented a small twin-jet business transport and was investigated with various The circular symbols show the variation of drag coefficient nacelle arrangements.

with Mach number for the basic wing-fuselage-empenn lage mounted pylons and nacelles increase the drag to symbols. The calculated skin friction of the pylon nacelle combination when added to the baseline configuration wouldgive a drag variation CIS indicated by the top line of the cross hatching. The installati,on penalty for the propulsion system package i s small i n this case and there i s some beneficial interference at the highest speeds. Even lower values of the complete aircraft dmg can be obtained by the alterations to the nacelle installation as indicated on the right side, where decreases in drag due to changing longitudinal location, increasing incidence angle and cant angle and changes i n the effective area distribution in the fuselage nacelle region are shown.

References Effects of Aft-Fuselage-Mounted NacelI es on the Subsonic Longitudinal Aerodynamic Characteristics of a Twin-Turbojet Airplane. Lawrence E. Putnam and Charles E. Trescot, Jr. NASA TND-3781, December 1966.

I -

o

I

n

0 0

t

- 382 in the Langley l6-foot transonic tunnel. The mo model testing by splitting the fuselage a f t of the the vehicle suitable for a Shuttle Traini shows flow through na A s reverse thrust power level increased and speed decr got progressively closer and in some cases enveloped the horizontal tail. This resulted in a dec ase in tail contribution to stability and control effect as well as increased trim changes and horizontal tail dynamics. Reverse thrust plume visualization was obtained by injecting water into the exhaust and verified the force data. Exhaust plume impingement can be a problem with conventiona!

aircraft a t high angles of attack.

Jet Effects on Aero Characteristics t interference effeck on comp ete aircraft aerodynamics testing techniques for complete powered models are usually more difficult because of support system and inlet simulation problems.)

This single-engine four-jet V/STOL type aircraft was tested with an injection propulsion system. The exhaust nozzles, two on each side are located close beneath the wing. The data represent the change in aerodynamic coefficients caused by a change from power-off to power-on flight. Results are shown as a function of angle of attack for M = 0.8 and horizontal tail deflections of 0 ’ and

5”. In this case, jet effects are not large. Jet operation decreased I ift and

drag and increased pitching moment. When referred to the absolute values of the coefficients required i n flight, these increments represent a reduction in lift and drag respectively, of about 5 and 10 percent. AI though the magnitude of pitching moment coefficient increased due to simulated jet operation, only slight changes in model longitudinal stability were found, I have attempted t o illustrate some problems that may be associated with the inegration of the power plant into the airframe. The examples illustrated are probably much moie severe than those occurring with general aviation aircraft.

NASA has the experimental facilities and is developing the analytical tools which will aid. in the reduction of interference drag and provide guides for the best ways to incorporate the power plant into the aircraft.

a

o *

c

IrL

a

a

I -

" &

et

' 0

..L

0 c3

a

Fi

Dm/Df= b. 1.08 X;/D, =.I75

L,/D,= 1.51 D,,/D, = .a84

D~~/D,,, = .?35

20" Fan Pressure Ratio 1 .15 Nacelle .

Figure 2.

Design drag divergence Mach number: 0.8 Tightly cowled: D d D f = 1.08 Inlet: D ~ ~ / D ~ = .935 Inlet capture.mass flow ratio: Ao/Amax*)des, = 0.66 k- 75 Inlet cowl length: X./D, = .I75 Fan Boattail angle: f 6 O \ Q: .02- %

cc

\ h

a

-\ -

w

t Y

w

-

.01 v , w Variation of Inlet and Fan Boattail Pressure Drag with Mach Number Figure 3.

1. Below drag rise the inlet pressure drag was less than the estimated value. However, in the same speed range boattail drag was higher than the value estimated from model boattail tests.

2. In addition, as the inlet went into drag rise (above M=0.8) the boattail drag decreased somewhat.

3. Both of the above trends indicate the possibility of an interaction between the inlet and aft end flow fields for close coupled propulsion systems like this one.

!NUT PRESSURE

I #- P O T E I a t A L FLOW ANALYSIS

DRAG cow,

* DPI

' % h A x

I t I f

.4 0 6 0 8 LO t 2

INtn-BOATTAIL PROXIMITY, XlDEilw(

Effect of Boattail Proximity on lnlet Pressure Drag.

Figure 4.

To verify the interaction between inlet and exit flow fields, we analyzed the effect of the proximity between the inlet and nozzle on the inlet pressure drag using a 2D potential flow program. This was done by varying the distance between the inlet and boattail. The calculated pressure force wos adjusted to pass through the experimental value shown at X/D = .9.

As the boattail was moved closer to the inlet (decreasingX/bma )yo"

it resulted i n a reduction of the inlet drag; thus indicating that h e r e i s an interaction between these two flow fields.

In light of this interaction between the inlet and aft end flow fields, i t may be quite important to simulate the proper flow fields of both components simultaneously when doing isolated propulsion system and propulsion system/ integration work. Three propulsion simulation techniques that are commonly used are:

1 . Flow thru nacelle--which i s normally used to properly model the

inlet flow field.

2. The blown nacelle --for proper simulation of nozzle flow only (correct NPR) .

3. Powered turbofan simulator--close simulation of both inlet and nozzle flow fields.

At NASA Lewis we have a program underway to evaluate and compare these three simulation techniques (both isolated and installed with the airframe} in terms of their degree of simulation and their relative accuracy.

This program will be conducted for both convent7onal and unconventional types of airframe installation.

\

I - - - - -

+-t’iT+, - -

c=f-

Effect of Nacelle Size on DC-9-30 Cruise Drag.

Figure 5.

Flow-thru nacelles are normally adequate when concerned about the interaction between the inlet and wing flow fields like aft fuselage installations. Some of the results from the DC-9 refan program are i s the drag penalty associated with the larger shown above. Shown refan nacelle poltted as a function of free-stream Mach number, Mo(at CL = 0 . 3 5 ) . An estimate was made of the drag penalty and i s shown as a dashed line. Based on these wind tunnel results, the drag increment decreased as Mo was increased.

At the cruise Mach number of 0.78, most of the estimated drag incre- ment was cancelled out due to a favorable interference effect. This favorable effect was associated with the larger refan inlet and its closer proximity to the wing. This effect most likely occurs because the positive pressures on the stream tube suppress the wing upper surface velocities, thereby moving the wing shock forward and reducing the Mach number a t the shock with subsequent reduction in wing compressibility drag.

This trend was observed from wing pressure data. (Reference: NASA CR-121219. ) Figure 7. Full Span CTOL Aft Fuselage Drag Shown above i s the total aft fuselage drag with and without nacelles installed. The nacelles. hadoNACA-l inlet cowl contours and relatively low boattail angles (8 to 10 ). The estimated friction drag (flat plate type calculation) is shown as a dashed line. The reference fuselage did not have nacelles; however, it had the same total area distribution as the fuselage with nacelles installed.

For the reference fuselage, the measured drag was quite close to the calculated skin friction level. With the nacelles installed, the incremental increase in the drag a t Mach numbers from 0.7 to 0.97 was approximately equal to the increase i n skin friction drag associ- ated with the larger wetted area nacelles installed case. This cornpari- son indicates that the pressure drag of the isolated nacelles was essentially cancelled out when the nacelles were installed with the airframe. We found that this favorable effect was quite sensitive to inlet cowl geo- metry. When a cowl with u more blunt contour than the NACA-1 was tesied, a relatively large adverse effect occurred at these speeds.

Figure 8. Over-the-Wing Half Span spanwise position.

2 . Four different nozzle designs.

ons i n local wing geometry in the region where the exhaust flow passes over the wing.

4. Supercritical wing.

Figure 9. 3-D Newmann Representation o f O W Model An extensive aerod namic design effort was done on this model.

The main analytica r tool used in this design was the 3-D Neumann

Lifting Potential flow program. Shown here i s a graphical re- presentation of how the model was paneled up for this program.

d

z

~ I ! 0 V r3 -1 J

s

= 0.75 - 0.9

Because of the energy shortage and high fuel prices i t i s highly desirable to reduce aircraft drag and improve propulsion system efficiency. Two new propulsion system concepts we have come up with a t the LeRC are the high speed turboprop and the ducted fan.

These concepts show a large potential for reducing energy con- sumption (10 to 25%) compared to the conventional high BPR turbo- fan, These potential improvements would be obtained through in- creased propulsive efficiency.

NASA Lewis has recently initiated a high speed (M = 0.8) turboprop aerodynamic technology program. The model show&% the figure will be used to do part of this work. An 84" air drive turbine w i l l be used to drive 30" diameter highly loaded propellers. This turbine i s capable of producing over 1000 hp. The propellers w i l l have eight blades and be designed for M = 0.8 cruise at 35,000 feet. They will be tested i n the Lewis 8 x 6 SWT. Increased propeller efficiencies may be achieved if tailored nacelle blockage shapes behind the propellers can be designed to suppress the Mach number i n the pro- peller pbne (and reduce propeller compressibility losses) without incurring high drag themselves. These blockage shapes will be investigated on this m o d e l .

These very high power loading propellers (SHP/ a t take

off) have significant swirl thrust losses (6 to 8Oo A* ProPat )=70 cruise. I t

may be possible to recover part of these losses using a second counter rotating propeller, fixed stators, or even the wing. These areas w i l l also be investigated i n this program. A simulated wing is shown mounted on the model. The integration of the high speed turboprop with the airframe w i l l be further investigated on a small half-span aircraft model .

I

!

I \ I I \ I \ The unconventional ducted fan concept shown in Figure 12 may offer some advantages compared to the high speed turboprop. Some o f these advantages tire: smaller fan diameter, and reduction i n swirl and tip losses. I n order to make this concept viable, i t must have a minimum size fan cowling as shown. Conventional size cowlings tive to fan diameter) that are utilized with exist es would have very high cruise drags. This drag any thrust improvement obtained by going to this very high bypass ratio concept. A short cowl requires a fixed area nozzle and a thin inlet. The fixed nozzle with a large exit area results i n a high fan flow at cruise where ram effects increase the nozzle pressure ratio, and low fan flow at takeoff where there is little ram recovery. This wide weight flow range reqwires a variable pitch fan, Also, the thin cowl requires small flow spillage and therefore high fan flow to avoid drag problems at cruise. In addition to the high cruise flow, the low fan pressure ratio (FPR- 1.08) would minimize the amount of internal flow convergence at the nozzle exit and would result i n a large exit stream tube size and short boattail length. A t takeoff, the low fan flow dictated by the fixed nozzle minimize the sharp lip inlet losses, but some separated flow would occur during static operation.

Another concern that would have t o be evaluated i s the aeroelastic stability of the fan blades and cowling during operation with separated flow. This separated region would diminish as forward speed in- creased. No appreciable amount of acoustic treatment can be utilized with this fan cowling due to i t s small size. Therefore, a relatively low tip speed fan with low inherent noise would have to be incorporated. This concept i s currently being analyzed for NASA by Pratt and Whitney and General Electric under two study contracts (NAS3-19121 and NAS3-19201).

Mississippi State University A flight test method to determine the level flight performance of propeller- driven aircraft is currently being investigated at Mississippi State University. By measuring the amount of power it takes to overcome a known increment of added drag to maintain steady state flight conditions, it may be possible to determine the overall drag and the propeller efficiency of a general aviation-type aircraft.

Ropeller efficiency, q , is defined as the ratio of thrust horsepower to

P brake horsepower, or TV = ‘p 550BHP Equating thrust to drag by the thrust inclination angle Y gives

*

where vP = cosy.

P If an increment of drag A D i s added and power is increased such that the airspeed remains cwstant, then

* * . (D + A D ) V

(3)

P -I- “p = 550(F$HP + ABHP)

Using a propulsive efficiency factor

* * *

and eliminating drag gives

* ADV

I:-

‘Ip 550[ (Bill’ + A E P ) E - BHP]

P Expressing BHP a s the product of torque, Q, and propeller rpm, n, and

*

in (2) gives the basic incremental drag equation substituting for q P If there i s little or no change in propeller efficiency, E =: 1 and P AD D = Q G Thus we have expressed the total drag of the aircraft as a function of three easily measured parameters. A standard propeller torquemeter w i l l be used to Q and AQ, while a load cell attached to a drag chute will measure AD.

measured B y also measuring Y , the propeller efficiency rl can be computed directly from P (2) * Equations (6) and (7), however, neglect changes in induced drag and profile drag. It i s expected that the profile drag will remain constant, but it may be necessary to consider ADi, the change in I ift-dependent drag. By using a parabolic polar for the aircraft and including an amount of lift AL, it is possible to express the drag as 2a 0 AaW I- T I AR e ADD I)= TI A R e [(l+ % ) E - 11 .

Q P where a . i s the slope of the lift curve, Acx i s the change in angle of ,attack, W i s the aircraft weight and ADD is the incremental drag of the parachute. Further., e is the OswaId efficiency factor and AR is the aspect ratio.

Note that this equation requires measurement of flight test variables and aircraft parameters that are not included in the simpler forms. If the added drag is small, it i s expected that equations (6) and (7) will be sufficient. However, the final form of the drag equation which will be most suitable remains to be determined.

The incremental drag method of determining aircraft performance appears to offer an excellent alternative to current flight test practice. The aircraft lift, drag and thrust values and propulsive efficiency are readily determined from a simple flight test procedure. It seems reasonabl'e to expect that sophisticated data acquisition and processing systems will be unnecessary since the flight test i s con- ducted under steady state conditions. The current practice of determining aircraft drag by the gliding flight method is tedious and provides no information concerning ~ e n t a ~ drag method appea ~ b ~ ~ ~ t y and utility o t re 9, I T 1 0 ENC E Possible Applications of Soaring Technology to Drag Reduction 9.1 in Powered General Aviation Aircraft J. H. McMasters and G. M. Pblmer, Pwtdue University 9.2 Minimum Vertical Tail Drag E . E. Larrabee, Massachusetts Institute of Technology

Preceding Page Blank

en era^ Aviation Aircraft John asters and George M. Palmer Purdue University Introduction The term "General Aviation" usually brings to mind the range of powered aircraft encompassing the Piper Cub through executive jet transport aircraft. Depending on one's definitions and biases, however, a case can be made for inclusion of other types of aerodynamically supported vehicles such as the sailplane and their powered derivatiyes (self-launched sailplanes o r motor gliders) and perhaps even the lowly hang glider. While participation in soaring in this country is rather limited and the economic impact of sailplane manufacture i s miniscule, the current level of technology in this

branch of I ight aviation is extraordinary - particularly in the areas of aerodynamic

efficiency and uti1 ization of advanced materials and fabrication techniques. The purpose of this brief discussion i s to outline the present state-of-the-art i n soaring performance and review some of the techniques (particularly in the area of drag re- duction) used to achieve this performance. It can legitimately be objected that the performance requirements of sailplanes and I ight powered aircraft are quite different and that sailplane manufactures are not bound by the same economic constraints a s their counter parts in powered flight. However, to ignore the aerodynamic lessons learned in sailplane development would be, in our view, a serious oversight. In view of the fact that sailplane technical literature i s infrequently consulted by many aeronautical engineers, particularly those at universities, this brief review i s considered appropriate.

State-of-the-Art Most modern souring aircraft are pure sporting devices, the most elegant and

advanced of which are optimized for competition - which today imp1ies racing. The

classic design problem i s one of optimizing an aircraft for two design points: (1) low speed (minimum sink rate) flight in a rectilinear or banked turning attitude to max- imize rate of climb and (2) minimum glide angle (or maximum lift-drag ratio) in high speed rectilinear cruise. In racing performance, however, absolute maximum L/D

Preceding Page Blank

g a "low" sink rate (e.g. 2m/sec) at the highest speed I * t present two major types of competition sailplanes are in wide spread use: Standard Class, with spans limited to 15m (49.2 ft) with water balast (to increase wing

-

loading in strong lift conditions) and only simple hinged flaps not connected to the ailerons permitted, and Open Class where anything is permitted. Under pressure mainly -_I_ from European designers, the Standard Class will be divided into two classes for inter- national competition after ? 976, with one branch becoming an "unlimited" class keeping only the 15m span limit and the other basically retaining the present Standard Class rules.

The other category of soaring device'of interest in this discussion, the "motor glider", is slowly becoming more popular in Europe and the United States. I t i s basically a moderate performance sailplane fitted with an engine providing i t with a self-launch and out-landing retrieval capability.

Some typical modern sailplanes and motor gliders are shown in Figure 1. Per- Performance formance and geometric data for several typical types are listed in Table 1 .

capabilities are further clarified in Figure 2. Also shown for comparison in Figure 2 are glide polars for several other types of low speed flying device from (1). There are few standard handbook type references available on sailplanes and soaring technology. Probably the best sources of information are Soaring magazine, Technical Soaring (12) and the pub1ications of the Organization Scientifique et Technique Inter-

-

nationale du Vol-a-Voile (OSTIV) available from the Soaring Society of America (SSA).

---I Important recent material i s available in (5.6).

The basic configuration of the high performance sailplane was well established prior to WWII. Performance increases since that time have been very large, however, due mainly to three factors:

1 . A greater appreciation o f the importance of the qual i t y of

the aerodynamic surfaces and the necessity of sealing gaps and flow leakage in reducing drag.

2. Advanced airfoil designs with greatly improved (compared with Gijttingen and NACA 4 and 5 digit airfoils) characteristic in the Reynolds number range characteristic of sailplane operation.

- -- - --

* The ideal sailplane glide polar would be as "flat" as possible over the widest possible

speed range. Sailplanes, as in the case of most other aircraft types, seldom ''cruise" at the speed for l/D max.

41 0 ction of fiberglass construction a ese f u c ~ o ~ w i l l be discussed in more detail later. Some comparison, based on data d pre-war technical vintage and modern technology are presented in Table 2.

41 1 a

/j

C,)..3 I I Wei he Ka 6CR

I

- 1

aw-

ASW 17 I Figure 1. Sailplanes and Motor Gliden 41 2 SIGMA.

41 3 !

L I ma i " Figure 1 . (continued) Sailplanes and Motor Gliders 41 4 *.

0 (r w w w U)NN E h W N vlui .vi u!

Y - a h N InN N N 0.r m o l cool 8 h m 2 - N ?

*r a h W X Y m d \ > -I- OD-W +-mN +, m t;

.-

c v)

.-

t c O N W oulu m-.- )r m

e

f

W m

-

\ 0 O h O h h c c I ? - I I U c Y c L I 0 all-. UIO Y % h h u- L U w w '3 K X K I L L 4 X U - c u l W W- . .

(Dah N'G mu) N N - N N N N u .

r z -LL LL =v h w,

f

-i 0 Y E h u

-

. n m c 2 [ : n N N

-

- u

w u L z In n U n V z a a a U V a a

3 W

3 c c LI.

a UJ

V 9 '

a - c U . e i Y ai Ii 0 - 0 0 0 0 0 0 0 0 0 ua O O a l O N M P I I I I I I I 1 I - O 0 ) L U n h c E c C > w C w (r a w (v J a I -

f

z CI) C

.-

n U N

-

a I E < 41 6 STD C I R R U S N7889 OFLIGHT TEST RLSUllS WING D R A G B A S E D 0 I Y 5 400- v) t; w 500- + a a 600 - 700 - COMPLETELY\ TURBULENT WING \ 8 0 0 . -

’\

I I I I 1 1 I 30 40 50 60 70 8 0 90 100 110 120 1: D FLIGHT SPEED . MPH I

i

Figure Z.(continued) Still Air Sea Level Glide Polars for Several Natural And Man-Made Flying Devices 41 7 le 2, Wei he ASW-12 (1938) (1964)

Min. 5 120 fpm 109 fpm i- 9%

V a t Zmin 35 k t 43 k t -23% 43.3 +38% 31.5 '-/%ax 48 kt +17% V a t L/Dmax 41 kt

V at i = 6 fps 62.5 kt 88 k t +41%

t/D at i = 6 fps 17.6 24.7

Ka 6 CR Std. C i r r u s % Improve.

( 1955) (1969)

[din. 2 134 fpm 134 fpm 0%

V at Zmin 36 kt 42.5 kt -1 8% 29 35.2 +21% L/Dmax

v a t L/Dma, 42 kt 51 kt +21%

v at i = 2m/s 70 kt 85.5 kt 4 . 2 2 %

21.9

L / D a t i = 2m/s 1 8

It should be noted that these gains in aerodynamic efficiency have not been accompanied by serious deterioration in stabil ity, control or safety.

Technical Considerations A number of practical factors make the sailplane design problem difficult, not a l l o f which are directly related to the absence of an engine. For good climb performance (low sink rate) a low wing loading, low weight and excellent aerodynamic efficiency ore desired. Further, if climbing i s to be done predominantly in thermals, trim drag in moderately steep turns must be low and the speed for minimum sink rate (maximum climb rate) should be low to minimize turn radius (which, for a given bank angle, varies directly with speed squared). On the other hand, for high speed cruise the main concern is aerodynamic efficiency (high L/D). In a first order analysis (i.e. neglecting Reynolds number effects), L/D is independent of weight and thus for a given wing area, a "high" wing loading is desired. The obvious solution of use of variable geometry (e.g. Fowler flaps) t o ameliorate the wing loading conflict is limited by several factors (e.g. class 41 8 ufacturing difficulties) some of es, e ~ n o m ~ c and/or drag conside ands of high aerodynamic efficiency in both ise ~ e q u ~ r e hat great care be taken to inimize both parasite and in- duced drag. The latter is ''easi f accomplished by use of high aspect ratio wings of near ideal planform. Given presently achievable values of parasite drag coefficient (about 0.010 based on wing area), the optimum compromise aspect ratios for Standard Class (span limited) sailplanes are between 18 and 22. Corresponding values for Open * C l a s s machines are between 25 and 30.

The use of high aspect ratio wings of moderate area at typical sailplane speeds, means that the wing operates in a Reynolds number range well below that of conventional GA aircraft. For example, assuming a machine with an aspect ratio of 22 and wing

area of 110 ft , with a useful speed range of 40 to 100 kt, the corresponding Reynolds

number range (based on average wing chord) is 1 .O to 2.4 x 10 at sea level. If the machine weighed 700 Ibs. loaded, the corresponding lift coefficient range at sea level would be CL = 1 .18 to 0.19, The general speed/Rn ranges for several types of low speed flying machines are shown in Figure 3. Sailplane experience indicates that with a little care, GA aircraft designers need not be overly concerned about the adverse influence of lowered Reynolds numbers on wing drag when large reductions in wing chord are contemplated.

Parasite R a n Reduction Post-war advances i n sailplane performance began when Raspet (7) demonstrated the shrtling performance gains achievable by systematically cleaning up a machine of initially good aerodynamic layout. The machine used was Dick Johnson's one-of-a- kind RJ-5 Open Class sailplane which was of conventional layout and construction (largely

wood) employing an NACA 6-series laminar flow airfoil . The results of the successive

improvements resulting from careful sealing of gaps and leaks, and reduction of wing waviness and roughness are shown in the now classic Figure 4. The total cleanup resulted -. ... .- .. . ..-. ..

* The "optimum" in this case i s not really clear, although practice indicates that machines with span greater than about 22 meters encounter serious flight and ground handling problems.

Required wing area spends on the extent to which variable geometry can be achieved and desired wing loading Thus overall operational consideration definin span and area

timit optimum aspect ratios based on achievement of pure maximum L ? D in both Standard

and Open Class machines.

41 9 of which was ach ie STD. S E A L E V E L C O N D I T I O N S MACH NO.

> n W 1 0 w - a - v) -

-

L.

I

-

- I R E Y N O L D S N U M B E R Rn V T / p Mote: Reynolds number based on average wing chord Figure 3. FI ight Speed/Reynolds Number Range for Various Low-Speed Flying Devices 42 1 Figure 4 . Results of Drag Clean-up on the RJ-5 Sailplane he r e q u i r e ~ e n ~ o f low drag over a reasonably wide I ift coefficient range and y low values of operat~ngReynolds number make airfoil selection for sailplanes somewhat difficult. Pre-war sailplane designers rei ied primarily on Gzttingen and NACA 4 and 5 digit airfoil, some of the former type (e.g. G i j 549) being specifically tested for sailplane applications. The advent o f the NACA &series airfoils of substantial drag reduction over at least the CL range of the "la provided care was taken in manufacture, there appeared hope of obtaining the "bucket" in practice. A number of very successful designs were thus produced in the late 1940's

and 1950's using various NACA 6x - 4xx and 6x - 6xx airfoils; the moderate camber

of these sections representing a reasonable compromise for centering the 'I bucket" be- tween the high and low speed extremes in required C L ' The theoretical work of R. Eppler and F.X. Wortmann i n Germany, beginning in the 195O's, showed that by careful contouring of the airfoil envelope and camber line, the transition point on low-to-moderate speed laminar airfoils could be controlled with some precision. This work lead to a family o f Wortmann airfoils (the FX or Franz Xavier series) which have been almost universally adopted in sailplanes designed during the last decade. Wortmann's work is well summarized in his paper in (5) and his air- foils have produced something o f a revolution i n modern sailplane performance.

Wortmann has shown that by carefully contouring the wpper surface of a fairly highly cambered airfoil, the upper end of the laminar flow range can be extended to section C values required f o r low sink rate.

L When a highly cambered airfoil is operated at low CL values, however, the airfoil i s frequently flying at a negative geometric angle of attack, and thus the lower surface of +he airfoil is the one on which transition (and/or separation) is of primary Thus, by careful contouring of both the upper concern in maintaining low profile drag.

and lower surfaces, the low drag "bucket" can be significantly extended (in operating CL range) compared to NACA 6-series sections of similar thickness and minimum drag.

The extent of the bucket can often be further increased by adjusting the camber line

with a small chord (1 0 - 200/0) simple hinged flap at the trailing edge. Examples of the

possible improvement are shown in Figure 5. Several typical sailplane and related air- foils are shown in Figure 6 , and the general trend in maximum section lift-drag ratio with Reynolds number for several Wortmann and NACA Sections are shown in Figure 7 .

While Wortmann's results are impressive the limited data available on the new Liebeck (1 0) sections appears spectacular. Whether such airfoils, which appear to approach some sort of theoretical li ement airfoil lift-drag ratio, can perfor to a practical wing Wortmann's invest~gations of the same type airfoils is reported in his paper in (6).

High Lift Devices In the modern gospel of sailplanes airfoil design according to Dr. Wortmann the wing contour must be absolutely smooth and unbro leading edge high I ift devices, wide Fowler flaps of significant chord are out of the question in sailplane design. Thus the designers choice of high lift devices i s severely limited. As one example of a way to circumvent this problem, and provide the performance benefits theoretical I y available

from use of area changing flaps *, Wortmann tailored a unique airfoil/flap system speci-

fically for the very advanced British "Sigma" sailplane project (see Figure 1 ) .

The FX 67-VC-170/136 section for "Sigma" i s fully described in (1 1) and the combined polar at Rn=3xlO 6 is shown in Figure 5. The flap of this airfoil i s "hidden" inside the basic FX 67-VC-170 airfoil when retracted, thus avoiding flow disruption at high speed.

When extended, it produces a 36% increase in chord. An even more exotic scheme has been proposed and tested by Wortmann (1 1) which involves deploying a large sheet of sailcloth (e.g. dacron) allowing chord extensions of greater than 50% in the high CL range.

Structures The third component in the post-war revolution in sailplane performance has been the introduction of fiberglass a s the main construction material; as pioneered by Nggele, Eppler, Stender and Hade in Germany. The use of fiberglass wing skins allows fabrica- tion of relatively wave free surfaces of unexcelled smoothness. A further consequence of the use of relatively low modulus of elasticity fiberglass is that in order to maintain desired levels o f torsional and bending stifhess, wing skins must be quite thick and correspondingly stronger than required by existing sailplane airworthiness standards. One thus finds fiber- glass sailplanes with load factors approaching those of modern fighter aircraft with little weight penalty (due to the low specific gravity of the fiberglass). Considerable room for further improvements in structures exists by use of advanced composites and materials such as DuPont Kevlar (PRD-49) with nearly three times the modulus of elasticity of existing E-glass systems.

* Partial span Fowler flap systems have been extensively tested on the South African BJ- series of sailplanes with generally poor results.

Rn = 3 X 10 I I I I t a I * I I Section Profile Drag Coefficient Figure 5. Several Airfoil Drag Polars a t a Reynolds Number of 3 X 10 M A N - P O W E R E D A I R C R A F T SA1 L P L A N E S _I M A C A 63 -618 F X 0 5 - 1 9 1 mod.

3 F X 6 6 - S - 1 9 6

NACA 6 5 - 8 1 8 F X 67-K-I50 F X 63 -137 F X 6 7 - V C - I 7 0 N A C A 8418

/

FX 67-VC-170/136 Figure 6. Typical Sailplane and Related Airfoils rglass systems have severa disadvantages, however, Since most lanes are produced in limited quantities, most fabrication is done g in high cost and major qua ity control problems. Inspection for structural integrity remains a major difficulty. Questions also remain about the aging characteristics of existing fiberglass systems, and the problem of ul tra-violet degradation of the structure carries the commerical disadvantage of offering the customer a choice his aircraft. Thus, alternative fab tion schemes to pre- of the basic color scheme of serve the beneficial aerodynamic qual ities of fiberglass construction, while reducing cost, etc. continue to be explored. Notable among these schemes are the use of bonding (e.g. the Liaster If Nugget") (4), the use of "plastic" coatings over conventional alum- inum structure (e .g. Schweizer 1-35) (4) and the extrusion of major structural assembl ies as described by Morel I i (5).

Concluding Comments The design of modern high performance sailplanes is an enormously challenging task. The absence of an engine means that the designer cannot indulge in the "luxury" of simply fitting the machine with a larger powerplant to obscure deficiencies in aero- dynamic design, weight overruns or to provide the basic design with "growth potentiaf" .

Further, the machine must have very low drag values over a relatively wide lift co- efficient range and these values must be achieved in a relatively low Reynolds number range. It i s for these reasons that a careful study of the remarkably successful methods sailplane builders have used to achieve these goals may well repay the designers of powered General Aviation aircraft. It has not been the intention of the authors of this brief discussion to advocate or imply that all sailplane technology i s applicable to General Aviation aircraft in general or that the operational and economic constraints faced by the GA designer make incorporation of appl icable features feasible. However, even a brief examination of the performance figures achieved by modern soaring machines and a little reflection as the often huge disparity in L/D values between sailplanes and GA aircraft indicates that careful attention to the lessons learned in sailplane design and manufacture hold realistic promise for substantial gains in the aerodynamic efficiency of several GA types. The fuel crisis, whether transient or permanent, may force a re-

direction in GA design with greatly increased emphasis on operation "economy" , perhaps

even at the expense of speed (or productivity) and initial vehicle cost. Modern soaring technology indicates one path along which future development might progress.

for~ance of Wort~ann t y e sail plane airfoils in wings w i t h roughness, Ids numbers typical of light powered general av~atiQn air- craft need to be further investigated.

2. Simple, economical methods of construction need to he found which lead to improved surface finish and manufacturing tolerance to approach the performance of sailplane types far powered general aviation aircraft. Possible examples are:

-

Plastic" coatings over conventional structures.

(e.g. Schweizer 1-35 sailplane)

- Metal bonding techniques (e .g. the Laister 'I Nugget")

- Major assembiy extrusions from metal and plastic

The optimum configuration design implications of the use of sailplane technology 3.

in powered general aviation aircraft needs to be investigated. Specifically:

- optimum wing geometry

- engine/thrust producing device location

- requirements for high lift devices

- wingbody integration (i.e. wing position and body shape)

to minimize adverse interference 4. The economic and configuration imp1ications, in I ight of the fuel crisis and sailplane technology, of optimizing the design of a given general aviation air- craft toward maximum "transport economy" even at the possible expense of "productivity" needs to be evaluated.

References

1 . McMasten, J.H., "An Analytic Survey of Low-Speed Flying Devices -

Natural and Man-Made", AIAA Paper No. 74-1019, Sept.1974.

(Also contained in Ref. 6 below). An expanded version of this paper will be published in Tech. Soaring, Vol . 3, 1975.

-

Bikle, P. , "Polars of Eight - 1971", Soaring, June 197l 2 .

Zacher, H . , "Some Flight Tests on Self-Launching Sailplanes", NASA 3.

CR-2315, Nov. 1973. "Flug messung m i t 52 Segelflugzeugen", Aero Revue, Oct., Nov., Dec. 1973.

-

Soaring, August 1974 (Journal of the Soaring Society of America, P.O. Box 4.

66071 , Los Angeles, Cal ifornb 90066)

5, Nash-Webber, J.L. "Motorless Flight Research - 1972" NASA C R 2315, Nov. 1973.

6 .

I., 1974, (Ava~iable from SSA, P . 0 . 071 90066 for $1 0,) Raspet, A. "Systematic Improvement of the Drag Polar of the Sailplane RJ-5" , e Soaring, Sept., Oet. 1951.

8 . Wortmann, F.X,, 'I Drag Reduction in Sailplanes", Soaring, June and July 1 966.

9. Scheurnann, W. , 1972 Symposium on Competitive Soaring. (Available

from SSA).

Liebeck, R. H., "A Class of Airfoils Designed for High Lift in Incompressible 10.

Flow", AIAA Paper No. 73-86, Jan. 1973.

Wortmann, F. X. , "Airfoils of the Variable Geometry Concept" , OSTIV 11 .

Pub. XI, 1974. (available from SSA), Technical Soaring, Vol. I and I1 (4 issues each), available from SSA or 12.

D r . Bernard Paiewonsky, 9309 Burning Tree Rd. , Bethesda, Maryland 20034.

assac h use For an airplane with no asymmetric power problem (a glider or a cross- shafted twin) and with no directional stability constraint (a control configured vehicle, or a skillful pilot) there s t i l l exists a requirement for a vertical tail large enough to perform a coordinated turn reversal, that i s to maintain nearly zero side- slip when banking from a coordinated turn in one direction to a coordinated turn in the other direction, Resumably this vertical tail maneuver load requirement establishes a minimum tail sire and a corresponding minimum tail drag.

This i s explained by the aid of Figure 1. In a coordinated turn the angular velocity about the aircraft Z axis, R, i s proportional to the sine of the roll angle @ times the ratio of the acceleration of gravity g to the flight speed V.

Differentiation of the expression for R shows that the angular acceleration during a turn reversal i s proportional to the roll rate P = d+/dt, and i s a maximum as the aircraft rolls through wings level where cos $ = 1.

When the lateral control i s deflected to produce the roll rate P , it in- variably produces an adverse yaw due to rolling AN = $J2SbC, b m ( p ) P where Cn i s a dimensionless stability derivative depending primarily on wing P characteristics. For an elliptically loaded wing of moderate to high aspect ratio P b

the local lift vectors are rotated nearly through the local helix angle (2y/b) (m)

giving rise to a value of Cn approaching -CL/8, ,the negative sign indicating a P negative (adverse) yawing moment to be overcome by the vertical tail in phase with a positive rolling velocity. The adverse yaw due to ailerons themselves (CnaA) may actually make larger than (-c~/8)(%/2V) for a given value of P,’ but I retain the - C L / ~ result for simplicity.

When the vertical tail load i s multiplied by the tail arm to provide the yawing moment necessary to overcome the inertial resistance of the aircraft to yawing 1Z(dR/dt), and the adverse aerodynamic yawing moment due to rolling, it 43 1 s fou he which i s a convenient form since most lateral controls will produce a maximum value of the tip helix angle (Pb/2V) independent of speed if the control can be fully de- flected.

For example, let full aileron deflection produce a helix angle of 0.1 radian, let the vertical tail arm be 0.4 of the wing span, and let the radius of gyration in yaw be 0.25 o f the wing span. The vertical tail load in a turn i s then reversal = 0.0625 mg independent of the flight speed.

If the maximum lift coefficient of the vertical tail i s the same as the wing, a tail area o f 0.0625 times the wing area is then required to perform a coordinated turn reversal at minimum flying speed. It i s seen that an "STOL" conversion, which would double C L ~ ~ ~ for the wing relative to the vertical tail, requires doubling the tuil area. Short tail arm to wing span ratios ("tailless airplanes") are seen to require very large vertical tails for satisfactory lateral control. The desirability o f increasing fuselage length i n the interest of reducing vertical tail size is clearly seen,

I 1

R !

10. s

iscussion - Session I - Statws of Drag

10.

10.2 Discussion - Session I1 - Fuselage Drag

10.3 Discussion - Session 111 - Wing Drag

10.4 Discussion - Session IV - External Nacelle Drag and Interference