20th AIAA Applied Aerodynamics Conference AIAA 2002-2941

24-26 June 2002, St. Louis, Missouri

AERODYNAMIC PERFORMANCE ANALYSIS OF BODIES

WITH DIFFERENT CROSS-SECTIONS

Shadi Mahjoob*
Research Scientist, Aerospace Research Institute
Tehran, Iran

Mahmoud Mani†
Associate Professor, Aerospace Engineering Dept.
Amir-Kabir University of Tech., Tehran, Iran

Mohammad Taeibi-Rahni‡
Assistant Professor, Aerospace Engineering Dept.
Sharif University of Tech., Tehran, Iran

ABSTRACT
In order to optimize the geometry of the missile’s cross- CM Pitching Moment Coefficient
section for transportation purposes and also to obtain CN Normal Force Coefficient
higher aerodynamic efficiencies, non-circular bodies µeff Viscosity Coefficient
have gained substantial attention by many researchers. µ Molecular Viscosity Coefficient
In this work, we have compared the aerodynamic µT Turbulent Viscosity Coefficient
characteristics of two missiles having the same cross- ĸ Turbulent Kinetic Energy
sectional areas, but different shapes (one circular and ε Turbulent Dissipation Rate
one square with round corners). In order to differentiate ρ Density
the non-circularity and the fin effects, we have ρgi Gravitational Body Force
considered the bodies with no fins. A three- Fi External Body Forces
dimensional, compressible, stationary, viscous, K Thermal Conductivity
turbulent flow has been simulated using the FLUENT M Mach Number
CFD code with the standard k-ε model. The results AR Aspect Ratio
have been compared with the existing benchmark data b Diameter of Cross-Section
and with the data we obtained using DIGITAL r Corners’ Radius of Cross-Section
DATCOM and MISSILE DATCOM softwares. Our τ ij Stress Tensor
results indicate that, even though the square section
missile has more friction drag, it produces less overall INTRODUCTION
drag. Also, its lift is higher than that of the circular case Nowadays, one of the open discussions in the area of
and thus has a higher aerodynamic efficiency. missile aerodynamics is increasing the aerodynamic
Moreover, the rate of increase of the aerodynamic efficiency by optimizing the geometry of its cross-
efficiency with increasing of the angle of attack is section. In this regard, many researchers’ attentions
higher than that of circular section. have been towards the missiles with non-circular cross-
sections. Even though, this causes some inconvenience
NOMENCLATURE in the missile’s stability and control, with new advances
L/D Aerodynamic Efficiency in that area, the aerodynamists suggest non-circular
CL Lift Coefficient cross-sections without much concerns.
CD Drag Coefficient Jackson and Sawyer1 have experimentally
investigated bodies with elliptical cross-sections and
*
e-mail: shmahjoob@ari.ac.ir; at the time of this research, she noticed a considerable increase in aerodynamic
was a graduate student in Amir-Kabir University of Tech. efficiency (L/D) for horizontal elliptical cross-sections
†
Senior Member AIAA, e-mail: mani@cic.aku.ac.ir (compared with circular cross-sections). Also, the work
‡
e-mail: taeibi@sharif.edu of Graves2, which was performed for a wide range of
Mach numbers (from subsonic to supersonic) around

1
American Institute of Aeronautics and Astronautics
Copyright © 2002 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
elliptic and circular cross-sectional bodies showed cause reduction in the effects of the corner radii in the
similar results. Recently, Sharma3 investigated both value of CN for compound body-fin configurations.
computationally and experimentally a similar problem. However, to the same degree, the corner radii affects CN.
His results not only support the previous findings, but
also show increase in CN and CM (for horizontal GOVERNING EQUATIONS
elliptical cross-sections). In addition, they indicate that, The Reynolds Averaged governing equations include
in any angle of attack, by increasing ellipticity ratio, the continuity, momentum, and energy, which are in
aerodynamic efficiency is increased. general written as:
In family of missiles with non-circular cross-
sections, because of the storage and carriage purposes, ∂
the ones with square or rectangular cross-sections (with (ρ u i ) = 0
∂x i
round corners) are used more extensively, especially in ∂ τ ij
∂ ∂p
tactical applications. The no–guidance air-to-ground (ρ u i u j )= − + + ρg i + Fi
∂xi ∂xi ∂x j
missiles, which need to have the maximum possible
load density and the minimum possible drag, is one of ∂
(ρ u i h ) = ∂  K ∂ T  + u i ∂ p + τ ij ∂ u i ,
the typical missiles having non-circular cross- ∂x i ∂x i  ∂x i  ∂x i ∂x j
sections.1 where
In 1987, Sigal and Lapidot4 performed an extensive   ∂u i ∂u j  2
experimental investigation, in which three families of τ ij =  µ eff  +   − µ eff ∂ u l δ ij ;
  ∂x j ∂x i  3 ∂x l
 
missiles with same length and cross-sectional areas
were used. The shape of their cross-sections were µeff = µ + µT.
square (with rounded corners), rectangle (with rounded
corners), and circle. Their results showed that CN was The flow considered here is three-dimensional,
the highest for the horizontal rectangular case and the compressible, stationary, viscous, turbulent and single-
second highest for the square case. Of course, the phase. The fluid is air, for which the viscosity is
problem with the rectangular case is that, when the obtained using Sutherland relation. The free stream
missile rolls and its cross-section changes from Mach number is 0.8, the flow Reynolds number based
horizontal rectangle to vertical, its aerodynamic on missile’s length is 1.83×106, the static temperature is
efficiency drops drastically (even to less than the 268.305 °K, the total gauge pressure is 29066.31 Pa, the
circular case). Note, their results were consistent in both turbulence intensity is 10%, the characteristic length
cases of a missile with and without fins. scale (the missile’s diameter of cross-section) is 0.04 m,
Flow separation effects are highly related to the and the operating pressure is 55433.69 Pa.
corners of the square or rectangular cross-sections and
cause unfavorable aerodynamic instabilities.5 Many TURBULENCE MODELING
researchers have studied these corner effects, especially The standard κ-ε model has been used in this study. κ-ε
taking the radius of the round corners. These efforts model is a two-equation model, in which the averaged
have led to considerable decrease in the above Reynolds stress is proportional to the averaged velocity
unfavorable effects. Schnider6 showed that for the gradient and the proportionality constant (µT) is to be
square section, the increase in the corner radius found from the κ and the ε equations.8
decreases the overall CN and the roll angle effects.
Also, Daniel et al.7 studied the above radius COMPUTATIONAL METHODOLOGIES
experimentally using a square cross-section (with The FLUENT CFD code, which uses a cell centered
rounded corners) for different angles of attack. They finite volume method, has been used in this study. The
found that the effects of this radius, especially in angles implicit method implemented uses a segregated solution
of attack higher than 10°, are considerable. Note, CN method. Note, all the schemes used here are second
also decreases as the corners become more rounded. order. The SIMPLE algorithm with under-relaxation
According to their studies, in high angles of attack, for coefficients is used in the overall discretization of the
r/b=0.2, CN is close to the circular cross-section case equations. In order to reduce the dispersion errors (and
(r/b=0.5), while for r/b=0.1, CN is close to the square also to increase the speed of the computations), the
cross-section case (r/b=0.0). multi-grid approach has also been used.9
This test has also been performed for a case in which
four fins are attached to the end of the missile. For this GEOMETRIC MODELING
case, similar results have been obtained. Of course, since In this study, the aerodynamics of the bodies of two
the fins are the main devices for producing lift, they missiles, one having a square and the other having a

2
American Institute of Aeronautics and Astronautics
circular cross-section (with same cross-sectional areas)
have been studied (Fig. 1). Figure 2 shows the physical
domain used, which is cylindrical with a smaller radius
in the front. Note, due to the symmetric assumption
used, only half of the domain is considered here. Our
computational results indicate that this assumption has
not affected the results considerably. Note also that, the
outer boundaries of the computational domain have
been assumed to be 9 radii in the radial direction, 2
body lengths at the front and 3 body lengths behind the
missile. This size of the domain was shown to be
optimal using numerical experimentation.
COMPUTATIONAL GRID
A structured, body fitted, and non-uniform grid has
been used in this study (Fig. 3). A 169×21×21 grid
showed to result a grid independent solution.
BOUNDARY CONDITIONS
Different boundaries of the physical domain are shown
in Figs. 2 and 3. The wall of the missile (boundary A) is
assumed to be adiabatic. The free stream pressure is
assumed at the inlet and outlet of the domain
(boundaries B and C). The far field pressure condition
is used at boundary D. Finally, at the symmetric plane
E, the symmetric boundary condition has been
implemented.
DISCUSSION OF RESULTS
In order to investigate the effects of the shape of the
cross-sections, two different cases (circular body and
square body with round corners) were studied in this
work. The computations were performed for different
angles of attack, namely: 0.0, 2.06, 4.09, 8.18, and
16.27 degrees. The FLUENT CFD code, which has
been proven to work well for different flow regimes
around missiles 10,11 was used in this study.
The results are mainly in the form of aerodynamic
coefficients. The benchmark data used here are the
results obtained by Mani and Khajehfard12. Note, the
only difference between their and our missiles is that
they also have two thin fins connected to the end of
their missile. In order to solely investigate the body’s
cross-sectional shape effects (without the aerodynamic
influences of the fins), as mentioned in the introduction,
we considered the body-alone configuration. Since the
existing wind tunnel results include fins, it is not
possible to perform a direct comparison between our
computational and the wind tunnel results (especially at
non-zero angles of attack). This is why, for comparison
purposes, we also calculated the aerodynamic
characteristics using DIGITAL DATCOM and
MISSILE DATCOM softwares. Note, the reference
3
American Institute of Aeronautics and Astronautics
area throughout this work is assumed to be the missile’s
cross-sectional area.
First, using the DIGITAL DATCOM software, we
obtained the aerodynamic coefficients for the missiles
used by Mani and Khajehfard12. Figure 4 has compared
these results for the circular cross-section case. In
general, DATCOM results can have up to about 20%
error. Also, there are errors in the above experimental
measurements because of the lack of a dryer in their
wind tunnel, the model’s manufacturing errors, and the
tunnel being very old.12 Therefore, the 7.2% error in
drag coefficient and 18% error in lift coefficient of
Fig. 4 is acceptable and thus our DIGITAL DATCOM
results can be safely used for the verification of our
CFD results.
Actually, the above calculations were also
performed using MISSILE DATCOM software (Fig. 4).
The error was more this time (44% in drag coefficient
and 21% in lift coefficient). This may be because of the
fact that two relatively large fins of the missile have
made it look like an airplane.
Figure 5 compares the CFD results and the results
obtained using DIGITAL DATCOM software for the
body-alone missile used in this study. For this case, the
errors for CD and CL were 12% and 28%, respectively.
However, using MISSILE DATCOM, the related errors
were 45% and 10.5% for CD and CL, respectively. Our
CFD results, which have about the same amount of
error as the wind tunnel data, were somewhere between
the DIGITAL DATCOM and the MISSILE DATCOM
results, which can reassure our CFD results.
Comparison of Figs. 4 and 5 shows drastic effects of
the fins on the aerodynamic coefficients.
Since the fins used in the above mentioned
experimental work are thin and symmetric (at zero
angle of attack) the corresponding results can directly
be used for our CFD verification purposes. At M=0.8,
the fins’ drag is mainly due to friction. This is because,
at this Mach number, the wave drag is relatively small.
Also, since we have zero lift coefficient, the related
2
induced drag coefficient (~ CL ) is zero as well.
πeAR
Finally, since the fins are thin, their pressure drag is
also negligibly small. Thus, at zero angle of attack, the
only remaining drag is in the form of skin friction. Note
that, the friction drag of the body-alone configuration is
available from our CFD results. Now, since the surface
area of the missile’s body and the fins are known, we
can estimate the friction drag of the fins. Here, the
interference drag has been neglected and the authors
believe it is very small for the conditions used. If we
subtract the friction drag of the fins (which is almost
their total drag) from the wind tunnel result (the overall
drag of body-fin configuration), we can have the total
drag of the body-alone configuration, which can be
used as benchmark data for our CFD computations.
With regard to the assumptions used, this comparison
shows satisfactory agreements. Note, the zero-lift
result, obtained from our CFD computations, match the
wind tunnel results, as expected.
Figure 6 shows better aerodynamic performance of
the squared section missile (compared with the circular
one) at different angles of attack. This is because we
have less drag and more lift in the square case. As the
angle of attack is increased, the lift and drag
coefficients increase as well, as expected. Interestingly
enough, as the angle of attack increases, the results of
the two cases (circular and square) deviate more,
resulting in even better aerodynamic efficiency in the
square case (Fig. 6c).
Another point is that, the friction drag is about 30%
of the overall drag in both circular and square cases. As
expected, this drag is more for the square case due to its
larger surface area. However, note that the overall drag
of the square case is smaller than the circular case.
In Fig. 7, the aerodynamic results obtained from
DIGITAL DATCOM software for body-alone
configuration are shown. As expected, this figure also
shows less drag for the square case. However, the
results related to the lift coefficient do not follow the
ones obtained by experiment and by CFD. This could
be because, there is about 28% error in the related
DIGITAL DATCOM lift coefficient calculations.
To better understand the aerodynamics of the
problem, we can study the related flow physics. Figure 8
shows the static pressure contours at zero angle of
attack. This figure shows the stagnation point at the
front side of the missile. Towards the end of the missile,
however, we note a considerable amount of increase in
velocity and decrease in pressure. The pressure
difference between the front and the back (pressure
drag) is more in the circular missile (Figs. 8 and 9).
Also, Figs. 10a and 10b show maximum and minimum
pressures for the two circular and square missiles at
different angles of attack. From these figures, note that
the circular missile has higher maximum and lower
minimum at all different angles of attack. In other
words, the circular missile has higher pressure
differences at all those angles of attack (Fig. 10c).
Figures 8 and 9 show that, the pressure starts to
decrease once the flow passes through the missile’s
nose, as expected. We also note the symmetry in the
results related to the zero angle of attack cases, which
results in zero lift.
Figure 11 shows that as the angle of attack is
increased, the stagnation point and therefore the high
pressure region moves to the missiles’ lower surface.
Also, the decrease in pressure at the top surface and
4
American Institute of Aeronautics and Astronautics
thus the reason for the increase in lift with angle of
attack is clear from this figure. Note, the pressure
difference region between the missile’s top and bottom
is greater for the square missile (compare to the circular
one). This verifies the previous finding that, the lift is
higher in the square case.
REFERENCES
1
Jackson, C.M. and Sawyer, W.C., “Bodies with Non-
Circular Cross-Sections and Bank-to-Turn Missiles”,
Progress in Astronautics and Aeronautics, Vol. 141,
pp. 365-389, 1991.
2
Graves, E.B., “Aerodynamic Characteristics of a Mono-
Planar Missile Concept with Bodies of Circular and
Elliptical Cross-Sections”, NASA TM-74079, Dec. 1977.
3
Sharma, R.K., “Experimental Aerodynamic
Characteristics of Elliptical Bodies with Variation in
Ellipticity Ratio”, AIAA-2000-4505, 18th AIAA Applied
Aerodynamics Conference & Exhibit., Aug. 2000.
4
Sigal, A. and Lapidot, E., “The Aerodynamic
Characteristics of Configurations Having Bodies with
Square, Rectangular, and Circular Cross-Sections at a
Mach Number of 0.75”, AIAA, Inc., 1987.
5
Nielsen, J.N., “Problems Associated with the
Aerodynamic Design of Missile Shapes”, Proceedings
of the Second Symposium on Numerical and Physical
Aspects of Aerodynamic Flows, Long Beach, CA,
Jan. 1983.
6
Schneider, W., “Experimental Investigation of Bodies
with Non-circular Cross-Section in Compressible
Flow”, AGARD-CP-336, Symposium on Missile
Aerodynamics, Trondheim, Norway, pp. 19-1-19-15,
Sept. 1982.
7
Daniel, D.C., Yechout, T.R., and Zollars, G.J., “Experimental
Aerodynamic Characteristics of Missiles with Square
Cross-Sections”, Journal of S&R, Vol. 19, 1982.
8
Keimasi, M.R. and Taeibi-Rahni, M., “Numerical
Simulation of Jets in a Cross-Flow Using Different
Turbulence Models”, AIAA Journal, vol. 39, Dec. 2001.
9
“FLUENT 5 User’s Guide”, FLUENT Incorporated,
July 1998.
10
Birch, T.J., Wrisdale, I.E., and Prince, S.A., “CFD
Predictions of Missile Flow-Fields”, AIAA 2000-4211,
18th AIAA Applied Aerodynamics Conference &
Exhibit., Aug. 2000.
11
Bulbeck, C.J., Morgan, J., and Fairlie, B.D., “RANS
Computations of High-Incidence Missile Flow Using
Hybrid Meshes”, AIAA 2000-4209, 18th AIAA Applied
Aerodynamics Conference & Exhibit., Aug. 2000.
12
Mani, M. and Khajehfard, A., “Aerodynamic
Characteristics of Bodies with Square, Elliptic and Circular
Cross-Sections”, AIAA 2000-4015, 18th AIAA Applied
Aerodynamics Conference & Exhibit., Aug. 2000.
 Fig. 1. A missile fuselage with two different cross sections.

Fig. 2. Half of the physical domain.

E E

D
D

E E

b c

Fig. 3. Computational grid: a) the symmetric plane, b) the circular body with one plane at an axial location,
and c) the square body with one plane at an axial location.

5
American Institute of Aeronautics and Astronautics
 Circular section Circular Section
Fin-Body Fin-Body

3.5 12
3
10
2.5
8
2
Cd

Cl
1.5
4
1

0.5 2

0 0
0 5 10 15 20 0 5 10 15 20
Alpha(deg) Alpha(deg)

Wind Tunnel Datcom Missile Datcom Wind Tunnel Datcom Missile Datcom

a b
Fig. 4. Comparison of the aerodynamic coefficients for fin-body combination using wind tunnel, DIGITAL
DATCOM, and MISSILE DATCOM: a) drag coefficient and b) lift coefficient.

Circular Section Circular Section

Body alone Body alone

3.5 12
3 10
2.5
8
2
Cd

6
Cl

1.5
1 4

0.5 2

0 0
0 5 10 15 20 0 5 10 15 20
Alpha(deg) Alpha(deg)

CFD Datcom Missile Datcom CFD Datcom Missile Datcom

a b
Fig. 5. Comparison of the aerodynamic coefficients for body-alone using CFD code, DIGITAL DATCOM, and
MISSILE DATCOM: a) drag coefficient and b) lift coefficient.

CFD Solution CFD Solution

0.3 1.2
0.29
1
0.28
0.8
0.27
Cd

0.26 0.6
Cl

0.25
0.4
0.24
0.2
0.23
0.22 0
0 5 10 15 20 0 5 10 15 20
Alpha(deg) Alpha(deg)

Square Section Circular Section Square Section Circular Section

a b
Fig. 6. Comparison of the aerodynamic coefficients and efficiency for square and circular missiles (CFD results):
a) drag coefficient, b) lift coefficient, and c) aerodynamic efficiency (c on the next page).

6
American Institute of Aeronautics and Astronautics
 CFD Solution

4.5
4
3.5
3
2.5

L/D
2
1.5
1
0.5
0
0 5 10 15 20
Alpha(deg)

Square Section Circular Section

Datcom Results Datcom Results

Body alone Body alone

0.6 1

0.5
0.8
0.4
0.6
Cd

0.3
Cl

0.4
0.2

0.1 0.2

0 0
0 5 10 15 20 0 5 10 15 20
Alpha(deg) Alpha(deg)

square circular square circular

a b
Fig. 7. Comparison of the aerodynamic coefficients for square and circular missiles (DIGITAL DATCOM results):
a) drag coefficient and b) lift coefficient.

Fig. 8. Static pressure contours around the square missile at zero angle of attack (CFD results).

7
American Institute of Aeronautics and Astronautics
 Fig. 9. Static pressure contours around the circular missile at zero angle of attack (CFD results).

CFD Solution CFD Solution

2.50E+04 0.00E+00

2.00E+04 -5.00E+03
Max Static Pressure

Min Static Pressure

1.50E+04 -1.00E+04

1.00E+04 -1.50E+04

5.00E+03 -2.00E+04

0.00E+00 -2.50E+04
0 5 10 15 20 0 5 10 15 20
Alpha(deg) Alpha(deg)

Square Section Circular Section Square Section Circular Section

a b

CFD Solution

4.50E+04
4.00E+04
3.50E+04
Pressure difference

3.00E+04
2.50E+04
2.00E+04
1.50E+04
1.00E+04
5.00E+03
Fig. 10. Comparison of the static pressure
0.00E+00
distribution for square and circular
0 5 10 15 20 missiles (CFD results): a) maximum
Alpha(deg) static pressure, b) minimum static
pressure, and c) pressure difference.
Square Section Circular Section

8
American Institute of Aeronautics and Astronautics
 a

Fig. 11. Comparison of the static pressure contours around square and circular missiles at eight degrees angle of
attack (CFD results): a) square and b) circular.

9
American Institute of Aeronautics and Astronautics