You are on page 1of 7

Minimizing

Fuselage Drag
BY BRUCE CARMICHAEL

INTRODUCTION ted area to the controlling frontal area. conventional fuselage. One enters on
3. Where practical, encourage ex- the horizontal scale, projects vertically
tensive laminar boundary layer flow. upward to the percent of length as-

N ow that wing designs are avail-


able with less than half the
profile drag coefficient of the
old favorite 23015 (Ref. 1), the fuse-
Concept (1) and (2) will be shown to
promote (3).
4. Fuselage shaping in the vicinity
of the wing trailing edge may possibly
sumed laminar line, then horizontally
across to the wetted area coefficient on
the vertical scale. Note the very large
drag coefficient reduction as transition
lage drag of even a relatively clean
airplane looms very large. The aircraft reduce fuselage pressure drag. moves aft from the nose to 60% of
designer finds considerably less drag length. You will note that the wetted
data for fuselages than are available for FORMULAS area coefficients for a given length RN
wings. This lack is true for bodies hav- and transition location are lower for
ing completely turbulent boundary Minimizing fuselage drag means the higher length/diameter case. How-
layers and even more so for those hav- minimizing the drag area which is the ever, later in our illustrative example
ing partial laminar flow. The reason product of the drag coefficient and the you will find that the drag area or prod-
for the latter is that the most popular area on which it is based. Calculation uct of the wetted area coefficient and
aircraft configurations have a multi- using the theory charts involves the wetted area will be lower for the low
tude of boundary layer trippers on the product of the wetted area drag coeffi- length/diameter case as will the frontal
front end including the tractor pro- cient and the wetted area. Most area drag coefficient. It is the drag area
peller. The subject of partially laminar designers favor presenting the result as that must be reduced. By the way, the
fuselages has thus been considered by product of frontal area drag coefficient length RN for standard sea level condi-
many to be academic. This article pre- and front area. The Young theory (Ref. tions can be calculated as 9354 x speed
sents theory and experimental results 2) gives wetted area drag coefficient as in mph x body length in ft. Another
for both fully turbulent bodies and ex- a function of: look at the magnitude of possible drag
tensively laminar bodies. A typical S = Wetted Area in sq. ft. reductions is given in Figures 4 and 5
fuselage drag minimization study is in- S = Frontal Area in sq. ft. for 1/d of 3.33 and 6.67 in the form of
cluded. The practical problems of 1 = Length in ft. ratio of drag with partial laminar flow
achieving extensive fuselage drag re- d = Effective Diameter in ft. divided by the drag for complete turbu-
ductions in practice are covered and Body length/diameter, length lent flow vs. the transition location.
the present state of achievement on ac- Reynolds number and fraction of Curves are presented for length RN of
tual aircraft is mentioned. Finally, a length laminar. The wetted area can be 1,10 and 100 million. It is possible to
performance estimate for an extreme estimated from the formula and Figure cut the drag more than in half.
application of fuselage drag reduction 1. Length RN = 9354 x speed in mph x
is presented. length in ft. ESTIMATING POSSIBLE
Kw = Wetted Area Coefficient EXTENT OF LAMINAR
FUSELAGE DRAG MINI- Cw = Wetted Area Drag Coefficient FLOW-IDEAL CASE
MIZATION CONCEPTS 05 = Frontal Area Drag. Coefficient
DA = Drag Area = CWS = C&S The boundary layer transition loca-
Total minimization by elimination tion for the ideal case of amplification
of the fuselage, while perhaps of inter- THEORETICAL of infinitesimal disturbances can be es-
est for gigantic thick wing, span loader, FUSELAGE DRAG timated by a complex computational
cargo or passenger craft, is not practi- method not available to many design-
cal for the modest sized manned The drag of a smooth, streamlined ers. I have presented my best guess for
aircraft of interest to us. Useful drag isolated fuselage can be rapidly esti- isolated smooth bodies in terms of the
reduction concepts include: mated by use of the charts of Young maximum transition length RN as a
1. Reduce fuselage surface area (Ref. 2). These charts are reproduced function of the length/diameter ratio
with pod and boom designs. here as Figures 2 and 3 for the case of (Figure 6). Dividing the transition
2. Use lowest possible pod, length to length/diameter = 3.33 typical of a length RN by the length RN then gives
diameter ratio, to reduce the ratio of wet- pod/boom design, and 6.67 typical of a the fraction of length ideally laminar.
64 AUGUST 1996
Symb i/d 10'6R^ CD (X/OTR Body

C5-"
r~i o K 3.320.019 0.45
U 2.5 0.30
52
0.019

0.0162 0.67 DolDhin


O 3.33 g5/250.013 0.60
(^ ^>>

A 5.0 4.4 0.022 0.70 sus

^ 3.0 0.037 0.82


</ 9.0 8.5 0.049 0.50 Parabolic
30 0.058 0.20
. . - • . • s'

v POD 5.5 7.0 0.034 0.34 ^_ ^ Altman


A p + B 10 10 0.031 0.34

TABLE 1 LOW DRAG BODY TEST DETAILS


You now know which of the Young seem to do as good a job as any. I ran shape parameter H or ratio of displace-
curves to stop at. The body design must such a study in 1964 where the appli- ment thickness to momentum thickness
promote flow acceleration (favorable cation was a high speed underwater at the most critical location on the af-
pressure gradient) to this fraction of towed sonar body. I used NACA 66 se- terbody was used to determine the
length to be valid. This curve is based ries low drag forms for the 30 inch degree by which dangerous separation
on experimental experience with diameter bodies. The study included was approached. A value of H = 2 is
smooth isolated bodies of NACA airfoil length/diameter ratios of 2, 2.5, 2.75, most often quoted as the danger point.
shape. One can perhaps exceed these 3.00, and 3.33. For each value, bound- The results of this study are plotted in
values slightly using more pointed ary layer transition locations of 0. Figure 7 on a grid of boundary layer
shapes as in "cross your feet, there is 20%, 40% and 59% were chosen. The transition position and body length/di-
only room for one rudder pedal."

ESTIMATING THE DANGER


OF AFTERBODY FLOW
SEPARATION
A look at theoretical drag predic-
tions tempts one to go to lowest
feasible length/diameter ratio and fur-
thest aft minimum pressure point. This
results in a very steep adverse pressure
gradient with danger of turbulent
boundary layer separation and very 3 4 5 6 7 8 9 10
large increases in drag. The prediction LENGTH-DIAMETER RATIO
of turbulent boundary layer separation
on non-conservative afterbodies is not FIG. 1 WETTED AREA CONSTANTS VS
an exact science. Some of the older
methods which track the variation in LENGTH-DIAMETER RATIO
the boundary layer shape parameter Kw = WETTED AREA / P1 (DIAMETER) (LENGTH)
SPORT AVIATION 65
ameter ratio. Boundaries separating
TABLE 2 DRAG MINIMIZATION CALCULATIONS safe, marginal, and separated regions
are shown by the hatched lines. The
Length Length Wetted Area KT6R, 10-6R60o/0 10-6Rtrans X/L trans. further aft that transition occurs, the
Diam. ft. sq-ft- _, lower the length/diameter ratio free of
separation. This is because the thicker
10 25 137 50 30 5.7 0.114 boundary layers with transition far for-
6.67 16.68 91.7 33.420 9.3 0.278 ward cannot surmount the adverse
5.0 12.5 68.7 25.0 15 12.6 0.504 pressure gradient as easily as when the
4.0 10.0 55.0 20.012 15.5 0.600 pressure recovery is stated with a thin-
3.33 8.33 45.8 16.710 18.0 0.600 ner layer. The length/diameter ratio 2
body was separated even when transi-
100%Turbuent tion was as far aft as 59%. The 2.5
Length Wetted Area Drag Coefficient Drag Area Frontal Area Drag Coeff. body was safe with 59% transition but
Diameter Sq Ft. separated with transition at 40%. The
3.0 body was safe at 59% transition,
10 0.00254 0.348 0.071 marginal at 40% transition, and sepa-
6.67 0.00280 0.257 0.052 rated at 20% transition location. The
5.0 0.00305 0.210 0.043 3.33 body was safe from separation
4.0 0.00335 0.184 0.037 even with transition at the nose and we
3.33 0.00372 0.170 0.035 were able to confirm this point on the
experimental model.
Partially Laminar '""* ' It should be remembered that these
Length X/Ltr. Wetted Area Drag. Coeff. Drag Area F.A. Drag Coeff. studies were for isolated bodies with
Diameter Sq. Ft. no wings intersecting. The data is
shown to provide a limit on how far
10 0.114 0.00226 0.310 0.063 one might go in combinations of
6.67 0.278 :;"""* 0.00257 0.236 0.048 length/diameter ratio and transition lo-
5.0 0.504 """" 0.00155 0.107 0.022 cation. Great care must be taken in the
4.0 0.600 0.00131 0.072 0.015 intersection region of low fineness ra-
3.33 0.600 "••••-- 0.00137 0.063 0.013 tio bodies and wings. This is a critical
area in sailplane design. There are now
computational methods powerful
enough to explore such configurations.
It would be well to search out those
.006 people operating such programs if a
serious design is considered.

FUSELAGE DRAG
.005
EXPERIMENTAL DATA
'f_

I had the good fortune to carry out


performance measurements on several
low drag bodies in low turbulence
wind tunnels, in a deep lake, and in the
ocean (Ref. 3). These bodies were sta-
bilized with boom mounted fins and
were driven by either positive or nega-
tive buoyancy. Frontal area drag
coefficients, presented in Figure 8, are
plotted vs. the body length/diameter
ratio and are compared to a similar
curve for fully turbulent bodies. The
three underwater partially laminar
bodies are supplemented with data I
obtained in flight test on a body with a
large value of length/diameter ratio,
12 5 10 20 50 100 and also some data on pod boom
sailplane body models. The body
LENGTH REYNOLDS NUMBER IN MILLIONS length/diameter ratio, length Reynolds
number, frontal area drag coefficient
FIG 2 STREAMLINED BODY THEORETICAL DRAG and transition locations are given in
66 AUGUST 1996
Table 1.
Designers are accustomed to frontal
area drag coefficients of 0.07 for .006
smooth but fully turbulent bodies with
length/diameter ratio of 6.5 as shown
on the upper curve. With a length/di-
ameter ratio of 3.33 and extensive .005s-
laminar flow, a frontal area drag coef-
ficient as low as 0.013 has been
measured or only 19% of the coeffi-
cient for a conventional turbulent flow
fuselage! Note that for a long slender
body with a length/diameter of 9, the
extent of laminar flow decreases and
the drag coefficient increases rapidly
with increasing Reynolds number due
to the very limited favorable pressure
gradient. At a length Reynolds number
of 3 million, laminar flow extends to
82% with a drag coefficient of 0.037
while at a length Reynolds number of
30 million, laminar flow extends to
only 20% with drag coefficient of
0.058. Low fineness ratio yields
stronger favorable pressure gradient 12 5 10 20 50 100
and retains extensive laminar flow to
higher Reynolds number. LENGTH REYNOLDS NUMBER IN MILLIONS
While working with Dr. Max
Kramer and designer Bill Knoll, I made FIG 3 STREAMLINED BODY THEORETICAL DRAG
the breakthrough experiment in 1961
with the Dolphin research vehicle my empirical curve of
1.0 ————
shown in Figure 9. The long slender Figure 6 to define the
bodies had been tested in the mid- maximum transition LEW ITH /I DIAM! TERI iATIO = 3.33
0.9 ^^\
1950s with disappointing results at the length Reynolds num-
higher Reynolds numbers. Our fat ber for each body \
body tests provide encouragement for length/diameter ratio. 0.8 ^i =-tS 46TH REYN DbOS NUMBER
INM LLIOt IS
certain extensively laminar fuselages The procedure is to h\

sK
in the Reynolds number range applica- calculate the length, 0.7
1100\
ble to aircraft. wetted area, length
Reynolds number for
PARTIAL LAMINAI DRAG
0.6 FULLY T IRBULEIi T DRAG \
\ ' ,'i

A SPECIFIC BODY DRAG each case. Maximum


transition length Rey- 0.5 ^
MINIMIZATION STUDY
nolds number from
\
\
0.4
Study the effect of body length/di- Figure 6 is divided by
ameter ratio on the drag of an isolated the length Reynolds
0.3
single person fuselage of 2 foot width number to get the frac-
and 3 foot height. This is representa- tion of body length 0.2
tive of single place sailplanes, racing laminar. The wetted area ) 0.1 0.2 0.3 0.4 0.5 0.6
planes and fast sportplanes. Equivalent drag coefficients are ob- TRANSITION LOCATION IN FRACTION OF BODY LENGTH
diameter is 2.5 feet and frontal area is tained at the proper
4.91 sq. ft. Assume NACA 66 series combination of length/ FIG 4 REDUCTION IN BODY DRAG
form with a wetted area constant of diameter, length Rey- DUE TO LAMINAR FLOW
0.7. Let speed at sea level be 214 mph nolds number, and frac-
which gives a unit Reynolds number of tion of body length laminar. These are decease linearly with decreasing
2 million per foot. Compute the frontal multiplied by the wetted area to obtain length/diameter ratio. At high
area drag coefficients and drag area for the drag area. The drag area divided by length/diameter, due to the higher
fully turbulent and partially laminar the frontal area gives the frontal area Reynolds number, the decease in drag
cases. The body design limits the max- drag coefficient (see Table 2). due to partial laminar flow is modest.
imum extent of laminar flow to 60% Frontal area drag coefficients are Below a value of 6 the reduction in-
which is the minimum pressure point plotted vs. length/diameter ratio in Fig- creases rapidly as we get into the range
but laminar extent may be less at the ure 10 and the drag area is plotted in where larger laminar length percent-
higher Reynolds number. We will use Figure 11. Fully turbulent body values ages are possible. Comparing a low
SPORT AVIATION 67
TRANSITION LENGTH REYNOLDS NUMBER IN MILLIONS
1.0
N LENC TH/[)IAME TERF ATIO = 6.67
0.9 ^S
vs
0.8 ^ LENC ITH R EYNOLDS
\ ^\NUMBEF 1 IN MILLIONS I
0.7 >
„««••

S
\>
\ \ \
100.
PARTIAL LAMINAF DRAG
0.6

g
1
FULLY Tt RBULEN 'DRAG
\
\ \

8
0.5
\ V. ^^

S
0.4
^-^

o
^

0.3

o
'• C 3 2 4 6 8 10
BODY LENGTH TO DIAMETER RATIO
0.2
C) 0.1 0.2 0.3 0.4 0.5 0.6
IRAN SITION LOCATION IN FRACTION OF BODY LENGTH FIG 6 EMPIRICAL ESTIMATION OF MAXI-
MUM TRANSITION LENGTH RN FOR
FIG 5 REDUCTION IN BODY DRAG DUE NATURAL LAMINAR FLOW BODIES AS
TO LAMINAR FLOW FUNCTION OF LENGTH / DIAMETER

drag pod design with a length/diameter without extension shaft and retain low placed where less laminar surface area
of 3.33 having extensive laminar flow, pod drag and good propulsion effi- is affected.
with a conventional fully turbulent ciency. Alex Strojnik built a very low A:'--' :• ' • • . I

fuselage with length/diameter ratio of drag airframe with such an arrange- OPERATIONAL
6.5, we find the drag area of the low ment (Ref. 6) and was very CONSIDERATIONS
drag case to be only a quarter of the disappointed in the achieved speed.
drag area of the conventional design. The most successful pusher aircraft in Just as work is the curse of the drink-
The pod design will require a boom to terms of propulsive efficiency have ing people, so insects are the curse of
support the tail surfaces. This would been those with the propeller behind laminar aircraft. Most insect impinge-
increase pod drag by 10% for a slender the tail. Molt Taylor worked out the ments will, especially at high speed at
boom up to 25% for a thicker boom. mechanical problems (Ref. 7) and Ed low altitude, trigger a turbulent wedge
With wing designs available with one- Lesher flew such a configuration, the with an included angle of 14 to 20 de-
third the drag of conventional turbulent Teal, to many world speed at range grees. They sometime erode with time
flow wings, the next step will be to records (Ref. 8). and at very high altitude can become
combine such wings with fuselage de- With the major disturbance out of subcritical and laminar flow can be re-
signs having the same small fraction of the way, we must now avoid the stored. The sailplane people have
conventional fuselage drag. While this smaller ones. Composite construction developed a bug wiper stowed at the
is not easy, it is possible. Composite can achieve surfaces free from wavi- wing root which is driven out along the
construction as in the wing case solves ness to a ratio of wave height to wave span by a vane and then reeled back in
many of the problems. length of one part in a thousand and with a wire. It actually clears off the
surface roughness to 10 microinch bugs and restores laminar flow. Such a
LAMINAR FUSELAGE DE- level. This is adequate for a small 300 device would be hard to develop for the
SIGN REQUIREMENTS mph airplane. There can be no protu- fuselage nose. Some success has been
berances, air leaks, gaps, grooves or found with liquid sprays to prevent bug
To obtain extensive laminar flow on steps in the forward fuselage. The adhesion until one climbs above the bug
the forward fuselage we must first re- fuselage back to the minimum pressure level. A disposable paper cover has been
move the propeller from the nose. Even point slightly aft of the maximum cross found workable on wing leading edges
though NASA has reported (Ref. 4) section must be made a single unit. if you do not get arrested for littering,
that the flow behind the propeller may This can slide forward on tracks for pi- but once again harder to work out for the
be less turbulent than previously be- lot entry and exit. This was proposed fuselage nose. The best way to have a
lieved due to the periodic nature of the in my articles in Sport Aviation in Au- laminar flight is to take off in early
prop wake, no comparative drag data gust and September 1976 (Ref. 8). morning before the bugs get up. Occa-
has been shown with and without pro- Cockpit ventilation should probably sionally at high altitude one encounters
peller. It is extremely difficult to mount not be attempted with an opening at ice crystals which, tumbling in the
an engine and propeller in the aft pod the nose. The air source should be boundary layer, shed eddies which cause
68 AUGUST 1996
DIAMETER = 30 INCHES
ZERO ANGLE OF ATTACK
NUMBERS REFER TO BOUNDARY LAYER SHAPE PARAMETER
AT MOST CRITICAL SPOT ON AFTERBODY 0.10
A,rbulent bodies
0.7 = 25x106
j/
/*
0.08
g 0.6
/
3 0.5
Drag
0.06
_. -I y/ f
aminar Bodies
0.4
a
qS ~ °
/
/+\ { = 3x10010
0.04 _ 52x10*
<*
*s
0.3 /*
A/
"H
0.02
^

0
§0.1 2 4 6 8 10 12
LENGTH/DIAMETER

2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 FIGS FRONTAL AREA DRAG
LENGTH / DIAMETER COEFFCIENTS OF PARTIALLY LAMINAR AND
FIG 7 LENGTH / DIAMETER- FULLY TURBULENT BODIES AS FUNCTION
TRANSITION LOCATION BOUNDARY OF BODY LENGTH /
FOR SEPARATION DIAMETER RATIO EXPERIMENTAL DATA
turbulence. This is quite rare. If you fly There are now a few powerplanes ABOUT THE AUTHOR
in rain, the flow will become turbulent with the propeller behind the tail
but laminar flow will be restored a few Bruce H. Carmichael, EAA 3133,
such as the American Cirrus and the
minutes after leaving the rain. Flight in 34795 Camino Capistrano, Capistrano
French Orion. It would be interesting
rough air can vary the angle of attack Beach, CA 92424, graduated with a
to check the extent of fuselage lami-
enough to increase wing drag. The fuse- degree in Aeronautical Engineering
nar flow with flow visualization.
lage should be less affected. from the University of Michigan in
PERFORMANCE { 1944. Among his teachers was EAA
APPLICATION TO ACTUAL member Ed Lesher. Bruce later worked
ESTIMATION at Chance Vought and Goodyear Air-
,
B
AIRCRAFT OF AN EXTREME craft in applied aerodynamics.
Modern production sailplanes enjoy
APPLICATION A chance meeting with Dr. August
some laminar flow on the forward I did performance estimations for a Raspet at a sailplane meet in 1949 led
fuselage and in general due to lack of very small record attempt airplane to to Bruce j o i n i n g him in boundary
propeller and with pod and boom fuse- meet the 661 pound takeoff weight layer control flight research using
lage configuration have somewhat limitation. Single place, propeller be- sailplanes. This in turn led to exten-
lower fuselage drag than powerplanes. hind tail, retractable landing gear. Span sion of his research of high subsonic
To date they have not gone to a sliding 14 ft., chord 2 ft., NASA NLF (1) 0414 speeds with Dr. Werner Pfenninger at
single unit forward fuselage design so F airfoil. AMW 90 hp 3 cyl. 2 cycle in- Northrop Aircraft, using an F-94A jet
the canopy joints no doubt trip the line engine. With extensive laminar aircraft. He later extended both nat-
laminar boundary layer. There have flow on wing, tail and forward fuse- ural laminar flow and artificial
been two sailplanes built with my pro- lage, top speed comes out 328 mph. boundary layer control for drag reduc-
posed unit construction. The Swiss That is really moving on 90 hp with tion with Dr. Max Kramer, using
Elfe seems to be very nicely done and 85% prop efficiency. If we lost all unmanned underwater vehicles.
Sutherland in Australia also used this fuselage laminar flow, the speed would Bruce has served on aerodynamic
feature on his MOBA sailplane. The drop to 294 mph, a loss of 34 mph. If committees for various soaring organi-
name stands for "My Own Bloody we lost all laminar flow the top speed zations, and written articles for soaring
Aircraft." He reports that the unit can would drop to 258 mph, a loss of 70 and aviation magazines.
be slid forward and back in flight with mph from the extensively laminar air-
very low forces. On initial opening the craft. With modern materials it should REFERENCES
negative pressure over most of it over- be possible to meet the weight limit 1. Carmichael, B.H. — The Ultimate
come the positive pressures at the and provide the required surface per- Laminar Airfoil? Kitplanes Nov. 1994
nose. Once the internal pressure be- fection. Not what one would want for a 2. Young, A.D. — The Calculation of the
comes negative this effect would of sportplane, but for a record attempt Total and Skin Friction Drags of Bod-
course reduce. At any rate, it appears and to promote interest in advanced ies of Revolution at Zero Incidence.
one could get out in an emergency. aircraft design, it is tempting. British R & M 1874 April 1939
SPORT AVIATION 69
FLY OUR
AWARD-WINNING
AIRCRAFT.

Dakota Hawk
FIG 9 PHOTO ILLUSTRATION OF DOLPHIN 1
AWARD-WINNING DESIGNS 3. Carmichael, B.H. — Underwater 5. Carmichael, B.H — Laminar Light-
• PROVEN Vehicle Drag Reduction Through plane, The Difficult Dream. Sport
Choice of Shape. June '66. AIAA Aviation Aug. and Sept. 1976.
Second Propulsion Joint Specialist 6. Strojnik, A. — Laminar Magic —
Conference. AIAA Paper 66-657. Sport Aviation, Jan. 1990.
4. Holmes, Obara and Yip — Natural 7. Taylor, M. — Shafts in Light Air-
Laminar Flow Experiments on craft — Sport Aviation, Dec. 1985.
Modern Airplane Surfaces. NASA 8. Lesher, E. — Teal Racer — Sport
Technical Paper 2256 June 1984. Aviation, March 1968. ^
RAG COEFFICIENT
§
o

• RUGGED

• STABLE
o
g
o
g

Super Koala
FRONTAL AR
o
o

• FACTORY SUPPORT AND

SERVICE SECOND TO NONE


O 2 4 6 8 10
• AFFORDABLE FLYING 10 FRONTAL AREA DRAG COEFFICIENT VS.
LENGTH/DIAMETER RATIO
'::."' .. ' -V j.. • • , - . ' . " • • • ' '.

FUSELAGE POD DRAG


WIDTH = 2 FT. HEIGHT = 3 FT. SPEED = 214 M.P.H.
FRONTAL AREA = 4.91 SQ. FT. NACA 66 SERIES FORM

.;ft
The Classic

flSHER Fir/MG PRODUCTS

P.O. Box 468-S95 • Edgeley, ND 58433


Phone (701) 493-2286 O 2 4 6 8 10
For full-color into pack on all 9 models, send K in U.S., $7 outside U.S. FIG 11 FUSELAGE DRAG AREA VS. LENGTH / DIAMETER RATIO
For information, use SPORT AVIATION'S Reader Service Card
70 AUGUST 1996

You might also like