Acta Polytechnica Vol. 44 No.

2/2004

Computational Fluid Dynamic

Simulation (CFD) and Experimental
Study on Wing-external Store
Aerodynamic Interference
of a Subsonic Fighter Aircraft
Tholudin Mat Lazim, Shabudin Mat, Huong Yu Saint
The main objective of the present work is to study the effect of an external store to a subsonic fighter aircraft. Generally most modern fighter
aircraft is designed with an external store installation. In this project a subsonic fighter aircraft model has been manufactured using
a computer numerical control machine for the purpose of studying the effect of the external store aerodynamic interference on the flow
around the aircraft wing. A computational fluid dynamic (CFD) and wind tunnel testing experiments have been carried out to ensure the
aerodynamic characteristic of the model then certified the aircraft will not facing any difficulties in stability and controllability. In the CFD
experiment, commercial CFD code is used to simulate the interference and aerodynamic characteristics of the model. Subsequently, the model
together with an external store was tested in a low speed wind tunnel with test section sized 0.45 m × 0.45 m. Result in the two-dimensional
pressure distribution obtained by both experiments are comparable. There is only 12% deviation in pressure distribution found in wind
tunnel testing compared to the result predicted by the CFD. The result shows that the effect of the external storage is only significant at the
lower surface of the wing and almost negligible at the upper surface of the wing. Aerodynamic interference is due to the external storage were
mostly evidence on a lower surface of the wing and almost negligible on the upper surface at low angle of attack. In addition, the area of
influence on the wing surface by store interference increased as the airspeed increase.

Keywords: computational fluid dynamic (CFD), wind tunnel testing, validation, aerodynamic interference.

installation. A generic model of one of the subsonic fighter

1 Introduction aircraft used by Royal Malaysian Air Force was chosen for
Fighter aircraft are mostly designed to carry stores such as the study. Wind tunnel testing and computational fluid dy-
launcher or external tank under the wing. When these stores namics (CFD) simulation were conducted to investigate these
are installed, the flow on its surrounding components such interference effects. The low speed wind tunnel with size
as the control surfaces can be considerably changed. This may 450 mm×450 mm was used to conduct the experiments and
introduce several aerodynamic interference characteristic commercial CFD software was used for the simulation. Other
such as changes in aerodynamics force, increase in turbulence milestones in this study include the verification and validation
and possibly flow separation, These phenomena may intro- process and the suitability of applying a commercial CFD
duce adverse effect on other aircraft components such as the code for predicting the wing and external store aerodynamics
horizontal tail and vertical stabilizer and consequently may interference effects.
affect the controllability and stability of the aircraft.
Research on external store installation is complex and
extensive. It covers several research areas such as aero- 2 Simulation and experimental works
dynamic, structure, flutter, physical integration, trajectory
The methodology adopted to conduct the study consists
prediction, aircraft performance, stability analysis and several
multiple engineering disciplines. However, the focus of this of a few steps. The first and foremost is to obtain digitized
work is to study the aerodynamic interference particularly on wing section geometry. The digitization process was done
the change in aerodynamic characteristics. The aerodynamic using Photomodeller software. Then the second step is to
characteristics are prerequisite for the other analysis, since construct a scale model of the wing based on the digitized
the aerodynamic data are required for a subsequent aircraft wing geometry using Numerical Control Machine (CNC).
structural analysis, stability analysis, performance analysis Then several series of experiment were carried out upon
and store trajectory analysis. The investigation of aerody- the scale model in the wind tunnel at a low speed which
namic characteristics in the external store clearance program approximately 22.8 m/s. The digitized wing geometry was
usually involved complex flow field study with multi compo- also used in the CFD simulation. Gambit preprocessor
nent interferences. Usually flow of such nature is investigated software was used to produce the necessary mesh. The
through wind tunnel testing besides the empirical methods. setup was then simulated using Fluent 5 CFD software and
The main objective in this study is to identify the interfer- the CFD simulation was carried out with various physical
ence effect of a subsonic fighter aircraft that is currently used models, numerical algorithms, discretization method and
by Royal Malaysian Air force with the present of external store boundary conditions. At the final step, the study was

© Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 9

Acta Polytechnica Vol. 44 No. 2/2004

Fig. 1: Photographs of the marked wing surface at the various projection

wrapped up by comparing both the computed and

measured results in investigating further the nature of
the interference effect of a wing and external storage
configuration.

3 Aircraft wing external geometry

digitization
Digitization of the wing geometry is vital in order to ob-
tain an adequate aircraft model geometry that represents
the real aircraft. A Photomodeler 3.0 software is used to
capture and digitize the aircraft wing external geometry.
The software captures the image of 84 photographs taken
at various angles of the aircraft wing. These photos were
taken using a digital camera. A number of points have
been marked on the aircraft wing and the adjacent fuse-
lage part using the masking tape as shown in Fig. 1. The
size of the markers was designed in such a way that the size
will be visible clearly and sharp in the photographs taken
at certain distance away. This was determined using the
Fig. 3: Digitized wing geometry after smoothed with CAD
relationship between the number of pixels and the dis-
tance from the camera. The placement and location of the
markers are determined based on the profile of the wing.
The high curvature area was placed with denser markers.
This figure also shown a total of 84 photographs were used
to generate the wing profile and some part of the fuselage.
Output from the digitization process is a set of coordi-
nates conforming to the wing geometry as shown in Fig. 2.
Majority of the coordinates were on the wing surface and
wingtip pylon. Unfortunately, the wing geometry image is
less quality in term of accuracy and perfection. Therefore,
CAD software is used to smooth the image. After made
minor adjustment, the image becomes as in Fig. 3.

3.1 Wind tunnel testing

A wing model is required to perform the wind tunnel
testing. Therefore, a 20 % scale wing model of a fighter
aircraft has been fabricated by using CNC machine. The
Fig. 2: Aircraft geometry produced by Photomodeller model was made from a single solid piece of an Aluminum-

10 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/

 Acta Polytechnica Vol. 44 No. 2/2004

-Alloy that having nine conduits each on the upper and lower storages. Meanwhile, the second configuration is with the ex-
surface. Fig. 4 show the semi-span model of a generic fighter ternal storages. Both configurations were tested at zero angle
aircraft taken from the digitized geometry produced by of attack at two different speeds, which are 22 m/s and 27 m/s.
Photomodeller.
3.2 Computational fluid dynamic simulation
In the CFD simulation, the mid wing was simulated at two
conditions. In the first condition, the mid wing have been
mesh into 111 239 elements. Meanwhile, and the second
condition it have been mesh into 221 112 elements as shown
in Fig. 6a and 6b. The flow was simulated at the speed 22 m/s,
incompressible flow and at laminar consideration. Fig. 6c
shows the simulation for wing with external storage with
122 158 elements.

Fig. 4: Semi span model of a generic fighter aircraft

The figure also indicates 3 main part of the wing which

includes a root section, mid section and tip section. But the
external storage is installed at the mid section. Then we only a) b)
decided to fabricate the mid section. Furthermore, we unable
to test the full set of the wing model due to the limitation of
the size of the wind tunnel test section. The model has three
main stations for pressure measurement study located at the
chord wise that parallel to each other with equal distance plac-
ing. Every station was stationed with static pressure-taping
point on upper and lower surface respectively. Besides built
the mid wing model, we also built a 1/5 scale model
of launcher and pylon as the external stores. These exter- c)
nal storages were design in such a way that they are easily
secured and removed from the wing section. Fig. 5 shows the Fig. 6: CFD model surface meshes:
complete assembly of this aircraft wing together with the a) mesh for wing in tunnel, 111 239 elements, b) mesh for
external storage, inside the test section. The experiments wing in tunnel, 221 112 elements, c) mesh for wing and
have been conducted using two different configurations of the store in tunnel, 122 158 elements
wing model. The first configuration is without the external

0.6

0.4

0.2

0
- Cp

- 0.2
coarse grid -lower
- 0.4 coarse grid - upper
fine grid - upper
- 0.6 fine grid - lower

- 0.8
- 0.2 0 0.2 0.4 0.6 0.8 1 1.2
x/c

Fig. 7: Pressure distributions at the mid span for upper and lower
Fig. 5: Model installation inside the wind tunnel surface

© Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 11

Acta Polytechnica Vol. 44 No. 2/2004

Pressure Distribution - st3 Pressure Distribution - st 1

0.6
0.6

0.4
0.4

0.2 0.2

0 0

store-st3-lower store-st1-lower
-0.2 -0.2
store-st3-upper store-st1-upper
-0.4 -0.4
clean-st1-lower
clean-st3-lower
-0.6 -0.6 clean-st1-upper
clean-st3-upper
-0.8 -0.8
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

X/C X/C

Pressure Distribution - st2

0.6

0.4

0.2

clean-st2-lower
-0.2

-0.4
clean-st2-upper

-0.6 store-st2-upper

-0.8
-0.1 0.1 0.3 0.5 0.7 0.9 1.1

X/C

Fig. 8: Pressure distribution for a various chord wise location

Pressure Distribution - st3 Pressure Distribution - st 1

0.6 0.6

0.4 0.4

0.2 0.2

0 0
store-st3-lower
-0.2 store-st1-lower
-0.2
store-st3-upper
store-st1-upper
-0.4
clean-st3-lower -0.4
clean-st1-lower
-0.6
clean-st3-upper -0.6 clean-st1-upper

-0.8
0 0.2 0.4 0.6 0.8 1 -0.8
0 0.2 0.4 0.6 0.8 1
X/C
X/C

Pressure Distribution - st2

0.6

0.4

0.2

clean-st2-lower
-0.2

-0.4 clean-st2-upper

-0.6 store-st2-upper

-0.8
-0.1 0.1 0.3 0.5 0.7 0.9 1.1

X/C

Fig. 9: Pressure coefficient at three different span wise locations

12 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/

 Acta Polytechnica Vol. 44 No. 2/2004

termining the angle of attack, accuracy of the model, block-

4 Results age effects and wind tunnel calibration can significantly
4.1 Wind tunnel testing results influenced the result. Even though the wing is machined
After conducted a series of experiment, the pressure dis- accurately by the computer numerical control (CNC), there is
tribution at mid span for the upper and lower surfaces was still doubt on the accuracy of the model. Moreover, the fluid
plotted as shown in Fig. 8. level of the manometer used for measuring pressure was fluc-
tuating between always from 1 to 2 cm.
From these results we found that at station 1, the differ-
ence in pressure coefficient is only 3 % on the upper surface of
the wing due to the external store installations. The pressure
5.2 Discussion
coefficient shows there are substantial differences by having In the study we observed that the external store configura-
external configuration at the lower surface. Station 2 and 3 tion only affects the lower surface of the wing. Fig. 10 shows
indicate the same phenomenon that there is a little difference the pressure distribution at the quarter chord point along the
in pressure distribution on the upper surface. The lower sur- span wise location from the tip to the root. At the upper
face shows some reduction in pressure distribution compared surface, the pressure distribution is almost constant from the
to the upper surface. These experimental results give an span wise position, which is from tip to the root of the wing.
initial indication that the flow at the upper surface will not be The pressure coefficient at the lower surface was reduced
severely affected by the external storage configuration com- by 40 % compared to the upper surface. With the external
pared to the lower surface. storage, the pressure distribution at the lower surface was in-
creased around 20 % compared with the clean lower surface,
4.2 Computational fluid dynamics results but there was sudden increased in pressure distribution at the
Fig. 9 shows the results of computational fluid dynamic position -0.2 span wise location where the external storage
simulation for the upper and lower surfaces of this aircraft. At was mounted to the wing.
station 3, it is found that the coefficient of pressure is almost
constant and same from the leading edges to trailing edge Conclusion
and the value is not much different by having the external
installation at the upper surface. This shows that the external The experimental study at speed 22.8 m/s and computa-
storage has not affected the flow at the upper surface. In tional fluids dynamic simulation has been performed on the
contrast the pressure coefficient is much differed at a lower wing and store configuration in this project. The results show
surface by having the external configuration compared to the that the flow over the upper surface of the wing has not
the clean wing configuration. This shows that the external affected much when the pylon and launcher are installed.
storage is only affecting the lower surface. The same phenom- The study also shows that the flow over the lower surface is
enon also happen at station 1 where the coefficient of much affected by the presence of external storage. The static
pressure has not changed at the upper surface but there are pressure around the wing is about 12 % higher than the
some reduction about 12 % in coefficient of in the lower simulated values.
surface. The simulation at station 2 also provides the same
results. References
Spanwise Pressure Distribution at Quarter Chord
[1] Manoj K. Bhardwaj, Rakesh K. Kapania, Reichenbach
0.7 E., Guru P. Guruswamy: “Computational Fluid Dynam-
0.6 ics/Computational Structural Dynamics Interaction
0.5 Methodology For Aircraft Wings”. AIAA Journal, Vol. 36
0.4
store - lower
clean - lower
(1998), No. 12, p. 2179–2185.
- CP

0.3
store - upper [2] Tomaro Robert F., Witzeman F. C., Strang W. Z.: “Simu-
clean -upper
0.2 lation Of Store Separation For The F/A-18c Using
0.1 Cobalt”. Journal of Aircraft, Vol. 37 (2000), No. 3,
0
p. 361–367.
-0.5 -0.4 -0.3 -0.2 -0.1 0
Span position (tip - root)
[3] Prewitt N. C., Belk D. M., Maple R. C.: “Multiple-Body
Trajectory Calculations Using the Beggar Code”. Journal
Fig. 10: External store interference on pressure distribution
of Aircraft, Vol. 36 (1999), No. 5, p. 802–808.
[4] Brock J. M., Jr, Jolly B. A.: Application of Computational
Fluid Dynamics at Eglin Air Force Base. 1988, SAE 985500.
5 Analysis and discussion [5] Shanker V., Malmuth N.: “Computational And Sim-
plified Analytical Treatment Of Transonic Wing/Fuse-
5.1 Comparison between CFD and lage/Pylon/Store Interaction”. Journal of Aircraft, Vol. 18
experimental works (1981), No. 8, p. 631–637.
The study shows that the value of pressure coefficient on [6] Hirsch: “Numerical Computation of Internal and Exter-
the upper surface predicted by simulation is around 0.4 com- nal Flows”. Volume 1. Brussels: Wiley, 1988.
pared to 0.6 performed by the experimental study. There is [7] Jameson A.: ”Re-Engineering the Design Process
about 12 % difference. In the experimental study, problem through Computation”. Journal of Aircraft, Vol. 36 (1999),
during the setup of experiment such as misalignment in de- No. 1.

© Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 13

Acta Polytechnica Vol. 44 No. 2/2004

[8] Spradley L. W., Lohner R., Chung T. J.: ”Generalized

Meshing Environment for Computational Mechanics”.
AIAA Journal, Vol. 36 (1996), No. 9, p. 1735–1737.
[9] Mavriplis D. J.: “Viscous Flow Analysis Using A Parallel
Unstructured Multigrid Solver”. AIAA Journal, Vol. 38
(2000), No. 11, p. 2067–2075.
[10] Kallinderis Y., Khawaja A., Mcmorris H.: “Hybrid Pris-
matic/Tetrahedral Grid Generation for Viscous Flows
around Complex Geometries”. AIAA Journal, Vol. 34
(1996), No. 2, p. 291–298.
[11] Koomullil R. P., Soni B. K.: ”Flow Simulation Using
Generalized Static and Dynamic Grids”. AIAA Journal,
Vol. 37 (1999), No. 12, p. 1551–1557.

Tholudin Mat Lazim, Ph.D

e-mail: [email protected]
Shabudin Mat, M.Sc
e-mail: [email protected]
Department Aeronautics & Automotive
Faculty Mechanical Engineering
Universiti Teknologi Malaysia
81310, UTM Skudai
Johor Malaysia
Huong Yu Saint, M.Eng
Royal Malaysian Airforce
Wisma Pertahanan
Jalan Padang Tembak
50634, Kuala Lumpur

14 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/