NASA

TECHNICAL

NOTE

NASA TN D-8260

O
(:O
!

A DESIGN

APPROACH

WIND-TUNNEL

AND

RESULTS

SUBSONIC

SPEEDS

MOUNTED

FOR

AT

SELECTED
HIGH

WING-TIP

WINGLETS

Richard

T.

lVhitcomb

Langley

Research

Center

_JJTIO/V

e_- A_,

Hampton,

Va.

23665
"2"76 .,L91 E,

NATIONAL AERONAUTICSAND SPACE ADMINISTRATION

WASHINGTON, D. C.

JULY 1976

1.

Report

No.

NASA
4.

Title

2.

Government

Accession

No.

3.

Recipient's

5.

Report

and

Subtitle

A DESIGN
RESULTS
MOUNTED

APPROACH
AND SELECTED
AT HIGH SUBSONIC
SPEEDS
WINGLETS

Richard

Performing

.Organization

Code

8.

Performing

Orgamzation

Report

NASA

Name

Langley

Hampton,

and

Work

Center

Agency

National

Name

15.

Supplementary

16.

Abstract

D.C.

'11.

Contract

Winglets,

Type

and

of

which

Space

Administration

faces

on the

extension

which

fuselage

juncture

(Suggested

are

forward
in the

forces,

transport
results

wing;

14.

vertical,
primary
the

addition

wing

of the

moments,
and

winglike
winglet

below
design

and
the

of the

for

of these

same

Sponsoring

Agency

Period

Covered

Code

report

above

presents

measured

effects

a representative
effects

reduce

rearward

drag

the

wing

a discussion
of these

sur-

first-generation,

with

in bending

those

for

moment

a wing-tip
at the

wing-

winglets.

Author(s))

18.

Distribution

Statement

Unclassified

Unlimited

reduction

drag
Subject

Security

and

Note

substantially
are

This

increase

systems

drag

tips.

loads

surfaces,

surfaces

winglets;

a comparison

in approximately

as did the

by

nearly
The

involved

lifting

Aerodynamic
Induced

small,

aerodynamic
jet

Nonplanar

are

conditions.

surfaces

considerations

narrow-body

19.

No.

Report

Technical

Address

at lifting

secondary

Words

Grant

20546

of the

Key

or

Notes

coefficients

17.

and

Aeronautics

Washington,

No.

23665
13.

Sponsoring

Unit

505-11-16-08

Address

Research

Va.

No.

L-10908
10.

Organization

Date

6.

T. Whitcomb

Performing

tips;

No.

July 1976

WIND-TUNNEL
FOR WING-TIP

Author(s)

12.

Catalog

TN D-8260

Classif.

(of

this

report)

20.

Unclassified

Security

Classif.

(of

this

page)

21.

Unclassified
* For

sale

by

the

National

Technical

No.

of

Pages

Category
22.

30
Information

Service,

Springfield,

Price*

$ 3.75
Virginia

22161

02

A DESIGN

APPROACH

HIGH

SUBSONIC

AND

SELECTED

SPEEDS

FOR

Richard
Langley

WIND-TUNNEL

WING-TIP

RESULTS

MOUNTED

AT

WINGLETS

T. Whitcomb

Research

Center

SUMMARY

Winglets, which are small, nearly vertical, winglike surfaces mounted

a wing, are intended to provide, for liftingconditions and subsonic Mach

at the tips of

numbers,

reduc-

tions in drag coefficient greater than those achieved by a simple wing-tip extension with
the same

structural weight penalty.

the tips. Smaller

secondary

The primary

surfaces may

be placed forward

includes a discussion of the considerations

measured

surfaces are located rearward

below the tips.

forces, moments,

the design conditions for a representative first-generation, narrow-body

in approximately

and loads near

jet transport

of these effects with those for a wing-tip extension which

the same

increase in bending moment

did the addition of the winglets.

transonic pressure

This paper

involved in the design of such surfaces; the

effects of these surfaces on the aerodynamic

wing; and a comparison

above

The experiments

at the wing-fuselage

were

conducted

results

juncture as

in the Langley

8-foot

tunnel.

For the configuration investigated the winglets reduce the induced drag by about 20
percent with a resulting increase in wing lift-drag ratio of roughly 9 percent for the design
Mach

number

of 0.78 and near the design liftcoefficient. This improvement

ratio is more

than twice as great as that achieved with the comparable

Also, the negative increments

in pitching-moment

of the w_inglets are less than those produced

wing-tip extension.

coefficients associated with the addition

by the wing-tip extension.

results also indicate that the increase in overall performance

the winglets in comparison

in lift-drag

The experimental

improvement

provided by

with that for a wing-tip extension is significantly dependent on

the angles of incidence of the upper winglet.

INTRODUCTION

It has
less

been

induced

Lanchester

recognized

drag
for

analyses

have

including

vertical

than

for

a planar

vertical
indicated
surfaces

many
wing.

surfaces
the

at the

years

that

As early
wing

tips.

significant

improvements

at the

(See

tip.

refs.

a nonplanar
as

1897
Since

lifting

a patent
that

possible
1 to 3, for

system

was

time
with
example.)

should

obtained

a number
nonplanar
On the

have

by
of theoretical
systems
basis

of

these encouraging theoretical studies, a number of experimental investigations of various

end plates at the wing tips have been made. Usually these surfaces have reduced the drag
at very high lift coefficients but have resulted, at best, in only slight reductions in drag
near cruise lift coefficients. Near cruise conditions the viscous drag increments associated with the end plates were nearly as great as the reductions in induced drag.
An analysis of the effect of vertical surfaces at the tip on overall airplane performance must include consideration of the effect of such surfaces on the structural weight.
Loads on the vertical surfaces and the increased loads on the outboard region of the wing
associated with adding these surfaces increase the bending moments imposed on the wing
structure. The greater bending moments, of course, require a heavier wing structure.
Aircraft designers have found that for the same structural weight penalty associated with
adding end plates, a significantly greater improvement in drag could be achieved by
merely extending the wing tip. As a result, no aircraft designs have incorporated such
surfaces for the sole purpose of reducing drag. However, vertical surfaces have been
placed near the tips of some sweptback and delta wings to provide directional stability.
The objective of the work described herein was to develop nearly vertical, tip mounted
surfaces which would provide, near cruise conditions, substantially greater reductions in
drag coefficient than those resulting from tip extensions with the same addedbending
moments imposed on the wing structure.
The factor that most previous experimental investigators have overlooked is that to
be fully effective the vertical surface at the tip must efficiently produce significant side
forces. These side forces are required to reduce the lift-induced inflow above the wing
tip or the outflow below the tip. Obviously, a low-aspect-ratio flat end plate as generally
tested previously is not an efficient lifting surface. To achieve the stated objective of the
present work, the nearly vertical surfaces placed at the tip for the purpose of reducing
drag due to lift have been designed to produce significant side forces, even at supercritical
conditions, according to the well-established principles for designing efficient wings; hence
the name winglets. Flow surveys behind the tip of a wing with and without winglets, presented in reference 4, indicate that the basic physical effect of the winglets, which leads
to drag reduction, is a vertical diffusion of the tip vortex flow at least just downstream of
the tip. The large inward components of the vortex flow near the center of the vortex are
substantially reduced while the small inward components in the region above the tip of the
winglet are increased slightly. Thus these surfaces could be called vortex diffusers.
The initial development investigations of wing-tip mounted winglets were conducted
at subsonic speeds on a representative second-generation, wide-body jet transport wing in
the Langley 8-foot transonic pressure tunnel during 1974. The results of that investigation
are discussed in general terms in reference 5. Complete results for the final configuration of that investigation are presented in reference 4. Recently, improved winglets have
been investigated on a representative first-generation, narrow-body jet transport wing for

a wider

range of flight conditions than that of reference 4.

approach

used in the design of the winglets, presents

recent wind-tunnel

investigation, and compares

This report describes

selected results from

the

the more

the results with the design objective.

SYMBOLS

The
to the
form

stability
on the

ficients
mean

longitudinal

for

axis
basis
the

of Units

and

exposed

wing.

and

in U.S.

used

semispan

of wing

local

chord,

cm

mean

aerodynamic

section

chord

chord

tip

Cb

bending-moment

CD

drag

chord

of exposed

lift

except

pitching

moments

dimensional

values

Units

reduced

(ref.

for

are

normal-force

All

in both

moment
Sb'

coefficient,

Drag

coefficient,

Lift
qoo S

of the
the

measurements

herein

are

defined

Interand

with

basic

tip,

wing

as follows:

124.26

semispan,

cm

(48.92

0.38h,

in.)

7.62

of exposed

basic

basic

wing,

39.98

S,

37.41

cm

wing,

obtained

from

cm

cm

(15.74

(14.73

integrated

in.)

pressure

wing

coefficient

coef-

quarter-chord

given

6).

referred

to coefficient

the

is the

are

Units.

coefficient

CL

been

report

(3.00

(in.)

of basic

in this

merit s

ct

Bending

have

wing,

Customary

in exposed

normal-force

me asure

All

data

basic

for

Customary

symbols

increase

Cn

U.S.

presented

moment

of the

of the

incremental

average

area
reference

Ab'

Cav

and

The

(SI)

made

Coefficients

exposed

chord

were

characteristics

Force

winglets.

System

calculations

system.
of the

aerodynamic

national

aerodynamic

of wing

at wing-fuselage

juncture,

in.)

in.

AC L

incremental

liftcoefficient at constant drag coefficient,

(CL)with

winglets
tip

Cm

or

-(CL)basic

wing

extension

pitching-moment

coefficient

about

moment

reference

Pitching moment

center,

qooS5

CN

normal-force

coefficient

on winglet

Cp

pressure

Cp,soni c

pressure

span

span

load

distribution,

based

PZ - Poo

coefficient corresponding

of upper

let

(see

inward

chord

fig.

plane

of wing

tip

(see

fig.

3),

20.32

cm

upper

from

free-stream

winglet,

with

direction,

leading

edge

positive

outward

for

with
lower

leadwing-

3), deg

Mach

static

measured
for

number

pressure,

Poo

free-stream

static

qoo

free-stream

dynamic

area

chordwise

spanwise

vertical

from

of winglet

free-stream

local

winglet

to local speed of sound

in.)

ing edge

integrating

qco

incidence

Pl

by

area

coefficient,

(8.00

Moo

obtained

of exposed

distance

distance

coordinate

N/m 2

pressure,

(psf)

N/m

pressure,

basic

wing,

from

from

N/m 2

0.4649

leading

(psf)

m 2 (5.0034

edge,

wing-fuselage

of airfoil,

(psf)

positive

positive

juncture,

upward,

ft 2)

aft,

cm

positive

cm

(in.)

(in.)

outboard,

cm

(in.)

z'

distance along winglet span from

OL

angle of attack, deg

DESIGN

chord plane of wing, cm

(in.)

CONSIDERATIONS

Methodology
The theoretical calculations of references

3 and 7 provide an indication of the span

load distributions required on the wing and vertical surfaces at the tip to obtain the optimum

induced drag in subcritical flow.

However,

they do not describe how the configura-

tion should be shaped to obtain these load distributions or how it should be designed to
achieve the maximum
the present

improvement

investigation were

in overall performance.

developed

The tip mounted

winglets of

with the available theoretical calculations, phys-

ical flow considerations, and extensive exploratory' experiments.

Consideration

has been

given to the effect of adding the winglets on the structural weight and the high-lift offdesign performance

as well as to the drag reduction at design conditions.

limitations of the methods

Because

used, the winglets of this investigation are undoubtedly

of the
not

optimum.
Since the development

of the design approach

ical liftingsurface methods

Among

ods should greatly aid in future aerodynamic

Calculations based on the method
made

of reference

3) is a nearly
is placed
are

vertical

reduced.

The

tion.

results
effects

probably

not

Conversely,

attachment

flow

of this

leading
to the

above
over

over

the

herein.
Some

winglets.

They have also

of the applications of

conditions,
ahead

the

of the

edge

of the

wing

becomes

configuration

the

wing

the

forward

The

surface

region

of the

design
that

to minimize

leading

edge

of the

winglet

a greater

crest
is moved
problem

1, 2, and

upper

winglet

of the

winglet

wing

conditions

suggest

upper-surface

upper

(figs.

tip.

inner

at supercritical

investigations

be significantly

surface

presented

winglet

velocities

effects

at supercritical

if the

of the

velocities

of exploratory

meth-

9 have already been used to verify a num-

rearward

increased

interference

These

Winglet

component

high

for

in this report.

mounted

the

on the

adverse

interference
should

so that

superimposed
Thus

primary

surface

rearward

not

face.

The

8 and 9.

nearly optimum

the design might be improved.

subsequently

liftingsystems

are references

designs of more

Upper
Arrangement.-

them

in the design procedure

indicated several areas where

this theory are described

herein, several new theoret-

for analyzing and optimizing nonplanar

subcritical flows have been developed.

ber of the assumptions

described

upper

sur-

are
adverse

root

of the

winglet

of the

wing-tip

sec-

aft
since

of this
the

crest,

structural

box

for the winglet moves aft of the usual rear spar location for the wing. Also, analyses and
exploratory experiments indicate that the shorter winglet root chord caused by moving the
leading edge aft of the wing section crest results in a perceptible loss of winglet effectiveness. Therefore the leading edge of the winglet has been placed near the crest for cruise
conditions. Results of exploratory experiments also indicate that the greatest winglet
effectiveness is achieved with the trailing edge of the winglet near the trailing edge of the
wing.
Loads.- The theories of references 3, 7, and 9 indicate that to achieve the reductions
in induced drag theoretically predicted for wing-tip mounted vertical surfaces requires
not only substantial inward normal loads on these surfaces but also significant increases
in the upward loads on the outboard region of the wing. Exploratory experiments made
both during the investigation of reference 4 and during the present investigation indicate
that the greatest measured reductions in drag due to adding the upper winglet are achieved
with normal loads on the winglet, and associated added loads on the outboard region of the
wing, substantially less than those indicated as optimum by the theories of references 3,
7, and 9. These differences are probably due primarily to viscous effects not included in
theory. Calculations based on reference 9 indicate that reducing these loads from the theoretical optimum values to the measured values decreases the effectiveness of the winglets only slightly (induced drag increases slightly). This effect is probably more than
offset by a reduction in viscous drag for both the winglet and the wing resulting from lower
induced velocities on these surfaces at the lower load condition. Further, with such
reduced loads the addedbending moments imposed on the wing and the resulting structural
weight increase are less than those associated with the theoretically optimum loads.
The theories of references 3 and 7 indicate that the optimum span load distributions
for the winglet are characterized by relatively high loads near the winglet root in comparison with the optimum elliptical load distribution for planar wings.
Height.- The available theories (ref. 3, for example) indicate that the reduction in
induced drag associated with tip mounted vertical surfaces increases slightly less than
linearly with increase in height. However, the theories indicate that the normal loads on
such surfaces and the loads on the outboard region of the wing required for the calculated
induced drag reduction also increase with an increase in winglet height. These greater
loads, together with a greater moment arm of the loads on the winglet associated with
increased height, of course, increase the bending moments in the wing with a resulting
weight penalty. Therefore, the optimum height must be a compromise between aerodynamic and structural weight considerations.
Further, the required normal-force coefficients for the winglet increase with an
increase in winglet height. For excessive winglet heights the required normal-force
coefficients would lead to substantial boundary-layer separation particularly for high-lift

off-design conditions. For the most satisfactory results the required normal-force coefficients for the winglet should probably be Iimited to values of the same order of magnitude as the lift coefficients of the wing.
The height of the wtnglet of the investigation described herein was selected arbitrarily on the basis of very limited exploratory experiments and anaiyses. A precise determination of the most satisfactory height must await more definite information on the
structural weight penalties associated with adding winglets.
Planform.- As for wings, the winglet should have the highest aerodynamic efficiency
when it is tapered so that the normal-force coefficient is approximately constant along the
span of the winglet. To achieve this situation for the desired span load distribution
requires substantial taper. For satisfactory winglet effectiveness at supercritical design
conditions, the effective sweep of these surfaces should be approximately the same as that
of the wing.
AirfoiI

section.-

requirements.
ficients

for

the

conditions,

upper

and

so that

delayed

to the

tive

should

devices

to the

cated

design

wing

coefficient

and

objective

the

on._the

NASA

general

that

without

causing

for

required

With

large

probably
airfoil

basic

normal-force

coef-

supercritical

design

a strong

wave

outboard

region

of the

Secondly,

the

separation

high-lift

for

the

described

wing.

provides

conditions

airfoil

on the

winglet

wing.

This

on the

Experiments

with

the

on the
wing

should

be

surface

is

Iatter

staI1

objec-

control

may

twist.since
to achieve

upper

the

effective
the

of camber
the

available

to chord

camber

analyses

have

high-speed

shouid

degrading

aerodynamic

simiiar

characteristics

the

significantly
and

be about

The

high-lift

an airfoil

a design

theoretical

to accomplish

or

with

10 with

and

thickness

structurai

ratio

effectively

low-speed

Also,

penalty

most
in reference

superior

weight

Because

winglet.

important

to avoid

on the

accomplished

a severe

amount

For

shaped

occurs

two

inward

number.

separation

low-speed

of airfoil

incidence,

of attack

such

ratio

and

of the

desired

boundary-layer

maximum

thickness

Incidence

substantial.

presence

Preliminary

satisfactory

the

the

characteristics.

characteristics.

geometric

are

an airfoil

the

velocities

to meet

wing.

that

supereritical

effectively,

induced

aviation

than

such

added

which

even

Mach

an airfoil

of significant
for

the

with

with

be achieved

objectives

factory

the

onset

These

lift

is achieved

conditions

greater

be shaped

provide

extended

icantly

should

efficiently

associated

shaped

airfoil

it should

to minimize

surface

winglet

First,

this

surface

The

be
the

studies

signifindi-

and

satis-

objectives

as low

as possible

low-speed

suggest

most

stall

that

the

most

8 percent.

winglet

is generaliy

inflow

desired
required
theories

angles

are

normal-force
in the

toed
greater

coefficient

winglet,

do not

out

as yet

this

and

thus

has

than

the

winglet

for

design

negative

incorporate

incidence
effects

negative
angles

conditions.
can be
of viscosity,

thickness, and supercritical flow, the most satisfactory incidence must at present be
determined by a systematic experimental investigation of various incidence angles.
To obtain the desired span load distribution on a swept upper winglet in an undistorted flow field would require substantial twist. However, the decrease in inflow with
increase in winglet height above the wing approximately provides the desired aerodynamic
twist. Thus, no geometric twist is usually required for this surface.
Cant
using

the

or

method

friction,

and

ration

dihedral.-

also

of reference

wing

should

bending

have

reduces

A study,

flow

on theoretical

7, of the

trade-offs

calculations
between

moments

indicates

amount

of outward

cant

at the

of the

a small

the

based

interference

that

root

the

made

induced

optimum

drag

Lundry
skin

winglet

in figure
winglet

L.

reduction,

practical

as shown
upper

by J.

1.

configu-

Outward

cant

at supercritical

conditions.

Lower
Rationale.as

one

than

of the

the

tions
tion

Theoretically,
same

height

optimum

height

of reference
with

drag.

improve

3 and

with

a resulting

mum

wing

edge

The

of the

winglet

this

most
ditions

satisfactory
is far

lower

winglet

nary

analyses,

contrast

to the

of ground
a lower

particularly
of a lower

of the

cases

of the

streamwise
of the

complete.

However

upper

winglet,

with

those
lower

camber

by calculations
the

lower

at the

the

leading

for

the

it is known

that,

(upper

surface

based

winglet,

as for

should

the

outward)

on reference
probably

surface.

in the

usual

maxi-

upper

winglet

at

near

the

is reduced.

trailing

edge
upper

high
the

complete

desired

edge

of the

induced

range
upper

9, also

It is

of the

root

velocities
of the

of flight
winglet,

toe-in.
suggest

be twisted

leading

winglet.

definition

and

super-

winglet,

inner

of a slat

of the

the

upper

of the

edge

lower

winglet

winglet

when

and

theoretically

reduction

interactions
on the

the

leading

may

coefficients

winglet

to that

optimum

pronounced

with

of the

similar

in

surface

on the

surface

in combina-

reductions

both

the

be shorter
calcula-

height,

lift

on the

inner

of flow

is nearly

configuration

substantiated

angle

effect

substantial

of the

high

velocities

maximizes

is roughly

local

wing

induced

must

a shortened

lessens

separation

region

effect

the

Because

requires

winglet

winglet

forward

such

is as effective

theoretical

additional

at both

optimum

lower

This

even

tip

usually

vertical

small

that

wing

The

of practical

relatively

indicate

the

problems.

winglet

in boundary-layer

coincides

from

clearance

produces

favorable

portion

tip.

measured

on the

Configuration.forward

winglet

presence

actually

in both

that

on the

a lower

that

coefficients.

conjectured
lower

However

effectiveness,

velocity

of a wing;

the

winglet

placement

lift

below

winglet,

decrease

induced

surface

experiments

7) and

Forward

vertical

However,

conditions.

(refs.

high

upper

overall

critical

a nearly

because

3 indicate

a larger

induced

above

Winglet

with

conthe

Prelimithat,

in

washout.

An analysis based on the theoretical results of reference 3 suggests that outward

cant of the lower winglet would increase the favorable effect of this surface on the flow
over the upper winglet. Therefore substantial outward cant was incorporated in this surface for the configuration of reference 4 and that of the present report. However, a more
recent analysis, using the method of reference 9, indicates that the h_ost satisfactory
overall performance is probably achieved with little or no cant in this surface.
EXPERIMENTS
Apparatus and Procedures
Test
pressure
The

facility.tunnel,

and

Mach

slots

allow

number,

A more

and

span

1.

was

narrow-body
the

wing

had

a slot

that

the

through

airplane

tions

the

for
are

at design

a Mach
presented

chord

line

lower

surface

band

of carborundum

sized
lets

upper

of the

on the
were

layer
forward

basis

located

conditions
to insure

was

winglets.

not

deflection
of

The

wind

range.
variable.

model

was

approximately

Reynolds
tubes,

tunnel

are

2 and

a semi-

shown
3.

num-

in fig-

The

basic

first-generation,
balance

midsection

but

covered

wing

was

the

Airfoil

did rotate

the

stiffness

balance

and

designed

same

C L = 0.44.

4.

inter-

speed

winglet

to the

The

in figure

wall

11.

in figures

and

section.

independently

possible

attached

Moo = 0.78

shown

subsonic

of a representative

protruded.

Boundary-layer

upper

and

of the

winglets,

These

strips

set

of reference

transition

are

shown

range.

strips.-

surface

13).

dewpoint

test

as that

so
for

the

The

model

tip

deflec-

coordinates

for

the

wing-

strips

were

I.

on the

rearward

the

in the

transonic

tunnel

measurement

model
are

fuselage

reduce

through

surface-pressure

The

are

section

highest

8-foot

rectangular

in reference

the

of the

bending

grains

(ref.

to obtain

those

transition
line

on the

is found

Langley

a slotted

test

and

approximate

of 0.78

in table

5-percent-chord

tunnel

conditions

number

of the

configurations

wing

tip

Boundary-layer
the

the

with

temperature,

angle-of-attack

which

in the

to be tested

to install

model

transport.

through

models

size

nacelles

nondimensional

actual

lets

jet

large

Photographs

of the
and

ceiling

In an effort

utilized.

wing,

and

of the

winglet

Drawings

fuselage,

with

description

sufficient

conducted
tunnel

floor

pressure,

description.-

model

ure

relatively

detailed

was

single-return

in the

stagnation

Model
ber

investigation

a continuous

longitudinal

ference

This

in a plastic
12.

The

in an attempt
The

lower

strips

ahead

surfaces
and

were

transition

shock

upper
wave

on the

on the

surface
the

surface

Reynolds
of the
various

(0.06-in.)

grains

lower

winglets
test

at

5-percentline

carborundum

full-scale

for

at the

placed

of a 0.15-cm-wide

The
strips

wing,

35-percent-chord

comprised

adhesive.

on the

of the

at the

to simulate

of the

transition

were
of the

number
were
conditions.

wing-

boundarylocated

Test
0.78

conditions.-

only.

kN/m

For

2 (850

per

Experimental

this

psf)

Mach

which

number

the

resulted

are

tests

trical

were

in a Reynolds

Force

strain-gage

was

measured

wing.

on the

with

moment

for

the

conducted

number

design

Mach

at a dynamic

of 17.2

within

the

number

of

pressure

106 per

of 41

meter

(5.25

106

were

the

Results
Results.of attack

The

with

for

the

and

a wing-tip

lift

basic

resulting

the

to changes

plete

full-scale

More

the

exposed

lift

coefficient

because

would

Span
are

coefficient
sented

root.

These

integrating
load

the
for

in figure

the

in figure

same
the

conditions

upper

10.

winglet

airplane

test

the

be less.

6.

(at a constant

juncture
were

obtained

for

the

distance

pressure

to the

Mach

various

presented
lift

not

balance

on the
and

coefficient

for

The
the

number.
than

for
in

estimated

that

full-scale

air-

at the

additional

surfaces

rollingby multiplying

center

to the

elastic

by the

balance,

were

the

ratio

total

associ-

increase

the

Selected

9.

a com-

total

wing

near

configurations.
in figure

for

those

be greater

the

measured

number

equiv-

determined

distributions

design
the

the

are

correcting

moments

from

increments,

by

changes

coefficient)

from

winglets,

penalties

the

angle

coefficient

been

lift

resulting

lower

relative

It has

in figure

and

Reynolds

would

shown

balance

are

the

those

by the

8 for

for

Therefore,

coefficient

the

one

drag

drag

than

shown

for

friction

for

measured

on

in figure

and

from

changes

side-force

distributions

differ

relative

vertical

and

presented

These

the

values

by the

stations

winglet

upper

6.

skin

those

would

wing-fuselage

The

the
than

are

factors
less

measured

presented
for

7.

increments
wing

by

butions

of the

upper

a constant

presented

complete

coefficient

in bending-moment

increments

obtained

cient

10 percent

axis

effects

the

for

in figure

configuration.

drag

of attack

coefficient,

winglet,

coefficient

less

for

on the

of 0.78

upper

conditions

to lift

compensating

in figure

of the

The

wind-tunnel

a constant

elastic

side-force

due

the

is presented

ratio.

angle

spanwise

pitching-moment
number

in lift

be somewhat

drag

of the

increase

is presented

axis

the

be about

horizontal

moment

would

two

with

surfaces

The

elec-

Discussion

Mach

At full-scale

additions

for

design

increase

lift-drag

airplane.

panel

The

the

in the

of these

plane

The

taken.

at four

stations

coefficient,

configurations

additional

importantly,

the

for

extension.

from

with

and

at the

a five-component

winglets.

and

of drag

coefficient

wing

alent

ated

variations

not

measured

at three

with

with

were

were

measured

configurations

obtained

fuselage.

distributions

they

the

were

measurements

located

static-pressure

for

data

Side-force

a device

In addition,

lower

and

balance.

Chordwise

10

presented

foot).
Measurements.-

the

data

and

winglets.

design

lift

pressure
of the

configuration

coeffidistri-

normal-force
is pre-

The effect of changing the incidence of the upper winglet on the drag for the configuration with both the upper and the lower winglets is shown
these incidence changes

in figure 11.

on the span load distributions are shown

The effects of

in figure 12.

Effect of upper winglet only.- Addition of the upper winglet only increases

the lift-

drag ratio of the exposed wing panel by about 9 percent near design conditions of
Moo = 0.78
and
ment
is decreased
strong

shock

results
lift

CL,basic
because

wave

indicate

in the

that

coefficients

the

value.
a vertical

is due

the

approximately

the

center

Adding

the

(fig.

5).

ficients
airplane,

effect

variations
approximately

and

the

shown

induced

drag

of this

values
values.

For

coefficients

ficient

for

the

winglet

is roughly

as the

design
3 for

winglet

9 indicate

that

based

investigated

the

.......

on ref-

also

slightly

associated

with

would

wing

(fig.

8).

be significantly

more

slightly

an

Calculations
less

for

a wing

the

winglet

more

design
curves
same

the

less

than

the

theoretical

on the

are

value
break

to the

ratios

lift

occur

at

the

the

0.7),
(fig.

the

minimum

of the

optimum

optimum

be somewhat
the

upper

configuration

of the

at which

winglet

for

two-thirds

may

coefficient

adding

For

one-half

(approximately

as the

in the

upper

values

9.

about

about

value

on the

with

3, 7, and

these

breaks

only

coefficient

loads

of references

are

the

complete

severe.

associated

wing

have

lift

coef-

the

would

added

with

wing

on the

for

coefficient

of the

winglet

pitching-moment

characteristics

Considerations,"

configurations,

the

negative

coefficient

are

region

loads

the

are

basic

With

but

loads

wing

pitching-moment-coefficient

span

7 and

investigation

airplane.

theories

added

from

winglet

in reference

Calculations

wing

aerodynamic

substantially

measured

other

predicted

outward.

pitching-moment

outboard

the

upper

up to the

to wing

in pitching-moment

"Design

by the

and

of the

coefficients

of height

in somewhat

coefficient

8, are

the

value

on the

on the

the

section

determined

At lift
and

in figure

the

of the

reductions

results

and

lift

on the

investigation

theoretical
oretical

of attack

in the

at higher

inboard.

for

loads

winglet,

the

trim

indicated
added

that
further

drag

lift

winglet

in this

distribution

winglet

same

Unpublished

effectiveness

of references

bending

load

changes

the

of the

obtained

thatthese

on the

tip.

winglet.

drag

located

ratio

theories

and

based

of angle

As

tilt

An analysis,

indicates

slight

to the

9 indicate

upper

the improveassociated
with

wing

addition

for

than

same

on the

span

of lift

the

based

of the

elliptical

on reference

with

with

in induced

and

in winglet

that

20 percent
greater

dihedral

effectiveness

Reductions

based

the

coefficients
separation

winglet

losses

5 indicates

by about

surface

that

of the
the

of figure

is substantially

primarily

9 indicate

juncture

numbers

lift

severe.

drag

Calculations

difference

increase

less

induced

mounted

investigated.

of the
Mach

results

reduction

tip

erence

much
of the

basic

This

region

at lower

are

An analysis
reduces

wing = 0.44
(fig. 6).
At higher
of wave drag and boundary-layer

the-

different.

angle-of-attack
normal-force

10).

The

coef-

decrease
11

in the relative magnitude

of winglet normal-force

coefficient at higher liftcoefficients is

associated with the unloading of the outboard region of the wing due to increased boundarylayer separation on the wing at these conditions.
Effect of adding lower winglet.- Near

design conditions of

Moo = 0.78

and

CL,basic wing = 0.44, adding the lower winglet has littleeffect on the lift-drag ratio
(fig.6). However, at higher liftcoefficients, adding the lower winglet results in a significant improvement

in the lift-drag ratio (fig.6).

As indicated in the section "Design

Con-

siderations," these favorable effects of adding the lower winglet for higher liftcoefficients
are associated with reductions in the relatively high induced velocities on the forward
region of the inner surface of the winglet and near the tip region of the wing opposite the
forward

part of the upper winglet (fig.9) with consequent

boundary-layer

reduction in shock-induced

separation.

Adding the lower winglet also reduces

and pitching-moment-coefficient
the break is about the same

the severity of breaks

in the angle-of-attack

curves (fig.5). With this surface added the severity of

as for the wing alone.

Adding this surface increases the loads on the outboard region of the wing (fig.8)
with a resulting increase in the bending-moment

increments

(fig.7). Data obtained at the one row of pressure

that, as for the upper winglet, the normal-force
the same

at the wing-fuselage

juncture

orifices on the lower winglet indicate

coefficients on the lower

winglet are about

as the liftcoefficients for the wing near design conditions.

An analysis of the effects of adding the lower winglet for all flight conditions indicates that the improvement
this surface, as suggested

in overall performance
in the section "Design

would be marginal.
Considerations,"

may

Modification of
change this

conclusion.
Effect of upper winglet incidence.- Near
C L = 0.48
let above

design conditions of

Moo = 0.78

and

for the configuration with winglets, increasing the incidence of the upper wing-4 (selected for the final configuration) increases the drag coefficient (fig.11).

This incidence increase also significantly increases

the loads on both the winglet and the

outboard region of the wing (fig.12) with a resulting increase in bending moments

imposed

on the structure.
Comparison
0.38h.

The added wing area is about 90 percent of the exposed

winglet.
same

with tip extension.- The tip extension investigated has a span equal to

The increase in root bending moment

as the increase for the upper

for the tip extension is approximately

design conditions of

wing = 0.44, this extension increases

Moo = 0.78

and

the lift-drag ratio by about 4 percent (fig.6)

which is less than half that achieved by adding the winglets.

12

the

and lower winglets, and it is greater than the increase

for the upper winglet only (fig.7). Near

CL,basic

surface area of the upper

Calculations of local bending moments along the span of the wing based on the span
load distributions (similar to those presented in ref. 4) indicate that adding the winglets
increases the bending moments on the outboard region of the wing by somewhat greater
amounts than does the tip extension selected. Studies made by industry have indicated
that such bending-moment differences usually have relatively small effects on the wing
structural weight.
The increase in lift coefficient for a given angle of attack associated with adding the
tip extension is about the same as that for the winglets. However, the increase in negative
pitching-moment coefficient associated with adding the tip extension is somewhat greater
than that for the winglets (fig. 5). With the tip extension the positive break in the variation of pitching-moment coefficient with lift coefficient occurs at the same lift coefficient
and has about the same magnitude as the break for the basic wing and for the configuration
with the upper and lower winglets.
CONC LU DING

A wind-tunnel

RE MARKS

investigation at high subsonic speeds of winglets mounted

a first-generation, narrow-body

jet transport wing has been conducted.

designed on the basis of the approach

presented

herein, are compared

extension producing

the same

as do the winglets.

Selected results are presented

1. At the design Mach

increase in bending moment

number

The winglets,
with a wing-tip

at the wing-fuselage

juncture

and indicate the following:

of 0.78 and near the design wing liftcoefficient of

about 0.44, adding the winglets reduces

the induced drag by about 20 percent and increases

the wing lift-drag ratio by approximately

is more

on the tip of

9 percent.

This improvement

in lift-drag ratio

than twice as great as that achieved by the wing-tip extension.

2. The negative increments

in the pitching-moment

the winglets are less than those produced

3. The normal-force

coefficient associated with adding

by the wing-tip extension.

coefficients for the winglets are about the same

as the liftcoef-

ficient for the wing near design conditions.

4. The magnitude

of the increase in overall performance

addition of the winglets in comparison

dependent

provided by

with that for the wing-tip extension is significantly

on the angles of incidence of the upper winglet and the associated loads on this

winglet and on the outboard region of the wing.

overall performance
minimum

improvement

The greatest measured

is obtained with substantially smaller

improvement

in

loads than those calculated for

induced drag by available theory.

13

5. An analysis
wing

tip

indicates

that

Langley

Research

National

Aeronautics

Hampton,
June

14

10,

Va.
1976

of the
the

effects

at all

improvement

flight

in overall

Center

23665

and

Space

conditions

Administration

of an auxiliary

performance

would

winglet
be marginal.

below

the

REFERENCES
1. Nagel, F.: Wings With End Plates.
Nov. 4, 1924.

Memo. Rep. 130, Eng. Div., McCook Field,

2. Mangler, W.: The Lift Distribution of Wings With End Plates.

NACA TM 856, 1938.

3. Weber, J.: Theoretical Load Distribution on a Wing With Vertical Plates.

No. 2960, British A.R.C., 1956.

R. & M.

4. Flechner, Stuart G.; Jacobs, Peter F.; and Whitcomb, Richard T.: A High Subsonic
SpeedWind-Tunnel Investigation of Winglets on a Representative Second-Generation
Jet Transport Wing. NASA TN D-8264, 1976.
5. Bower, Robert E Opportunities for Aerodynamic-Drag Reduction. NASA/University
Conference on Aeronautics. NASA SP-372, 1975, pp. 323-352.
6. Mechtly, E.A.: The International System of Units - Physical Constants and Conversion Factors (SecondRevision). NASA SP-7012, 1973.
7. Lundry, J. L.: A Numerical Solution for the Minimum Induced Drag, and the Corresponding Loading, of Nonplanar Wings. NASA CR-1218, 1968.
8. Lamar, John E.: A Vortex-Lattice Method for the Mean Camber Shapes of Trimmed
Noncoplanar Planforms With Minimum Vortex Drag. NASA TN D-8090, 1976.
9. Goldhammer, M.I.:
A Lifting Surface Theory for the Analysis of Nonplanar Lifting
Systems. AIAA Paper No. 76-16, Jan. 1976.
10. McGhee, Robert J.; Beasley William D.; and Somers, Dan M.: Low-Speed Aerodynamic Characteristics of a 13-Percent-Thick Airfoil Section Designed for General
Aviation Applications. NASA TM X-72697, 1975.
11. Schaefer, William T., Jr.: Characteristics of Major Active Wind Tunnels at the
Langley Research Center. NASA TM X-1130, 1965.
12. Braslow, Albert L.; and Knox, Eugene C.: Simplified Method for Determination of
Critical Height of Distributed Roughness Particles for Boundary-Layer Transition
at Mach Numbers From 0 to 5. NACA TN 4363, 1958.
13. Blackwell, James A., Jr.: Preliminary Study of Effects of Reynolds Number and
Boundary-Layer Transition Location on Shock-Induced Separation. NASA
TN D-5003, 1969.

15

'FABLE

I.-

AIRFOIL

COORDINATES

z/c
x /' c

for

WINGLETS

--

Lower
surface
0

0
.0020

0077

-.0032

.0050

0119

-.0041

.0125

0179

-.0060

.0250

0249

-.0077

.0375

0296

-.0090

.0500

0333

-.0100

.0750

0389

-.0118

.1000

0433

-.0132

.1250

0469

-.0144

1500

0499

-.0154

1750

0525

-.0161

2000

0547

-.0167

2500

0581

-.0175

3000

0605

-.0176

3500

0621

-.0174

4000

0628

-.0168

.4500

0627

-.0158

.5000

0618

-.0144

.5500

0599

-.0122

.5750

.0587

-.0106

.6000

.0572

-.0090

.6250

.0554

-.0071

.6500

.0533

-.0052

.6750

.0508

-.0033

.7000

.0481

-.0015

.7250

.0451

.0004

.7500

.0419

.0020

.7750

.0384

.0036

.8000

.0349

.0049

.8250

0311

0060

.8500

0270

0065

.8750

0228

0064

.9000

0184

0059

.9250

0138

0045

.9500

0089

.9750

0038

1.0000

16

Upper
surface

FOR

-.0020

0021
-.0013
-.0067

.........................
(a) Complete

L-75,8430

configuration.

L-75 -8429
(b)Winglets.
Figure

I.- Wind-tunnel

model.
17

O0

,/

,,

,.

,_....
._

zL___._---Z/---

i_k,_..J,,_

/ /

62.50(__.

t.6,,

.///
I

_
5 _. _(- _2_" ._2_
54

4 _(2!.44)

v
--7----7--12.80(5.04)

-'

, f"

\\
\

\ I
4\

\\\

'\

43.2i(i7.01)

\_\
p

!
84

0(33.

....... \X

_242t14892"

..........
Tip

extension

_as,c
t,_
I-2o.s
(8.oo)
7,62

275.09(108.30)
Spocer-_,

/--Simulated

_i_'_

half-fuselage

Moment

,_

reference

center

14._73(5.80)

Figure

2.-

Drawing

of semispan

model.

Dimensions

are

in centimeters

(inches).

(5.00)

t_-.21 ct_- _

Sect

Typical winglet section

Upper surface

Upper surface

Winglet
i,deg
Upper
- 4
Lower, root - 7
Lower, tip -I I

Span'h=ct

,J

///

/'

.65ct

I/
= ttt
r

O.
_.._52

-,.4

I_____---___i//

Upper surface

ct

_-.16ct
Ct

A/
Figure

_D

_ra

3.-

Winglets.

Tip configuration
Basic
Upper winglet
Upper + lower winglets
Tip

extension ,Ab'=0.58h

,10

J
J

j"
J
J

.06

Tip deflection
J

jj

..-_ j

J
J

.04

///-S

/
/.i

"/
J

J
J

..._

.O2

.I

.2

.3

.4
CL

Figure

4.- Wing-tip

deflections.

Moo = 0.78.

.5

.6

,7

,.._

o
o

O0

i
I

b_
,..-i

b_

"--"
e..l._l

4_

,,,_

o,-.i

b/?
]
i
N,,,-I

m
I

o ,.,,i

<D

04

0
I

_4

b_
,...i

-_
,,-i

Tip

[rl

configuration

Basic

[]

Upper

winglet

Upper

+ lower

Tip
:

winglets

extension,Ab'=0.58h
'i'"

.......

4
c_,deg
5

.04

.2

(b) Angle

of attack
Figure

22

and

pitching

5.- Concluded.

moment.

Tip configuration
Upper winglet
Upper + lower winglets
Tip extension,

.10

Ab' =0.58h

-.,.

.08

-.,,, \
\,

.06

AC L
-C L,basic

wing
.04

_j.

.02

L
0

Figure
b_

6.-

Variation

of

incremental

.I

lift

,2

coefficient

,5

for

constant

.4
CL,basic
drag

coefficient

.5
wing
with

.6

lift

coefficient.

.7

Moo = 0.78.

b_

Tip
Upper

configuration
winglet

Upper + lower winglets

Tip

extension,A

b'=0.58h

.O5

.04

._..,
fl
/
i
s

Cb

.05

CL

CL
basic

wing
.01

.I

.2

.5

.4

C L,basic
Figure 7.- Variation of incremental bending-moment

.5

.6

w ing

coefficient at wing-fuselage juncture

with liftcoefficient. Moo = 0.78.

,7

.8
Tip
Wing
.7

configuration

Basic

[]

Upper

Upper + lower winglets

z_

Tip

winglet

extension,

.6

Ab'=0.58h
I

.5

%
CnC
C

(:Iv

Upper winglet
.2

,2
CnC

Car I

,I

.I

.2

,4

.5

.5

,6

.7

.8

.9

1.0

bO

8.-

Span

load

distributions.

.I
z'/b'

y/b'
Figure

I.I

Moo = 0.78;

C L = 0.48.

.2

b_

Upper

surface
o

Lower

z_
-I.2

surface
[]

Winglet
On

Off

I
J

-.8

F Cp'snic

)
Cp

O
O

..... -_-_-_-:_
_ - - -o _- --

-. 4
D

5/x _

o
o

o
[3

ee

[]
O

[]

[]
13

y/b'--0.99

E3

z_/b':O.025

,4

-I.2

-.8
0
Cp -.4

oc

0
0

o
o
0

_P

E3 E3

[]

[]

F3

[]

[3

[]
[]

z_/b'=O.082
0

K_

[]

.2

.4

.6

.8

1.0

.2

x/c

Figure

9.-

.6
x/c

(a)
Chordwise

Configuration
pressure

[]

z_/b'=O.151

with

upper

distributions.

winglet.
Moo = 0.78;

C L = 0.48.

.8

Upper surface
0
z_

Lower

surface
D
<F

Winglets
On
Off

-I.2

-.8

6
A

_--

Cp -.4,

_ _

_,,,,
_08

.....

Cp,sonic
0

0
0

0
0

)0

0
[]

[]

q
[]

-iD

y/d--o.99
.4

[]

zpb'=O.025

[3

-I.2

-.8
O
-4

0 o

0
0

Cp
ODD

[q

[]

[]

[]

D
D

[]

z_'b'=O.082
.4

.2

.4

.6

.8

1.0

.4

.2

x/c

--.1

.6
x/c

(b) Configuration

with
Figure

9.-

0
D

zyb':O.ISl

upper

and

ConclUded.

lower

winglets.

.8

.0

_
,,-.d
I
I

O0

4
f

0
e,.l_l

o
D
cO

ho

t._

,.0
/

"-'_o

I--

o.
c'_

_.
Q.

I
/

I
I
I
I

.__I

LI
/

.,,_

"'_

(D

,-.,_

/,

0
_'_

.,_

0
0

0
0
,_

.,--_

0
o,,.i

Od

LO

o,J

""

-.2"
c-

d
O..
0..
r

28

._I
o

_-_

% _
o

o
o

p....
I

,,,.i

,--i

(D
O

I1

_N
CO
o

,-_

_J

0
\

\
o
o

o_

,....i

q-

to

o.

t_

o_
t_

PO

O0
c_

0
0

OJ
0

O_

OJ
O_

b/?
_

r.r-I
!

b.0

,-,i

29

,8
Upper winglet
incidence,deg

Wing
0
[]
0

.7

-5
- 5.5
-4

.6

.5

.4
CnC

Upper winglet
.2
CnC .2 _[
Cav

.I

0q

.I

,2

.3

.4

.5

,6

.8

.7

.9

1.0

z b'

y/b'
Figure

12.- Effect

of angle

of incidence

of upper

with both upper

and

winglet

on span

load

distributions

I
0
_0
0
CO

lower

winglets.

.I

Moo

= 0.78.

for configuration

>