“AEROSPATIAL 2010” – Section 1

Aerodynamic parametric study for a small UAV

Authors:
Mihai Victor Pricop, Eng., INCAS, Bucharest
Niculae Marin, PhD. INCAS Bucharest

Abstract: The paper presents results of a parametric study in an effort to develop a small UAV. The
design is constrained as payload, size, wing loading, endurance and many other certification
parameters, currently too early to be discussed. Preliminary vehicle configurations and their basic
aerodynamic characteristics are presented.

Requirements

The intention is to develop an imagery acquisition vehicle, provided with a gimbaled high definition
sensor. Some of the requirements for the presented vehicle are:

Ready to fly mass 3Kg

Payload 0.5Kg
Payload consumption: 10W
Endurance 2h
Hand launched
Belly landing recovery
Transportation box length: 700mm
Cruising speed 15m/s
Electric propulsion, noise level

Preliminary configurations

The transportation box size is a strong constraint, with respect to the wing span and it’s splitting in
panels (as well as for the fuselage). It is mandatory to assemble the wing from 0.7m panels. The
reasonable choices are three panel wings (two joints) and four panel wings (three joints). Each
joint has a weight contribution and decreases the reliability of the system (adding electric
connections). While keeping the wing loading constant (60g/sqdm) and increasing the span, we
have an increase in AR from 8.8 to 14, with a reduction of MAC from 0.237m to 0.187m, which
corresponds to a Reynolds number reduction from 0.23M to 0.18M. The airfoil performance is quite
sensitive with respect to Re, in this low Re regime, even if it is specifically designed.

The first configuration airfoil is KLH2411 from [1]. The author optimized the NACA 2411 using
XFOIL. The predicted performance of the optimized airfoil was 300% of the original, in terms of
maximum L/D ratio, which is not credible. The next three configurations are using a specially
designed low Re airfoil, SD6060 (Selig-Donovan), provided with reliable wind-tunnel data as in [2].
However, to better exploit this airfoil, transition tripping must be applied.

Config. Wing area[sqm] AR Wing airfoil VH MAC[m] Re[M] Remarks

1 0.48 8.8 KLH2411 0.263 0.237 0.23 Poor longitudinal
stability
2 0.48 8.8 SD6060 0.525 0.237 0.23
3 0.518 15 SD6060 0.523 0.187 0.18
4 0.515 8.57 SD6060 0.517 0.206 0.2
VH - volume of horizontal tail
MAC – mean aerodynamic chord
AR – wing aspect ratio

International Conference of Aerospace Sciences “AEROSPATIAL 2010” ISSN 2067-8622

Bucharest, 20-21 October, 2010 1
 “AEROSPATIAL 2010” – Section 1

Fig. 1 The four configurations

Starting with the second configuration, the fuselage is modified in order to accommodate a
clamping system able to release the wing in the case of a hard landing.
The CAD models for each of the configurations contain analytic defined curves for the airfoils,
obtained by constrained least square interpolation.

Fig. 2 Least square reconstructed airfoil - Fig. 3 Interpolation error

continuous curves; diamonds – original points;

f lower
( x)  y airfoil
The relative interpolation errors in Fig. 3 are computed as * 100% and
min( y airfoil )
f upper
( x)  y airfoil
* 100% .
max( y airfoil )
International Conference of Aerospace Sciences “AEROSPATIAL 2010” ISSN 2067-8622
Bucharest, 20-21 October, 2010 2
 “AEROSPATIAL 2010” – Section 1

The interpolation functions are f
upper coefficients and basis functions are:
6 9
lower
( x)   c i  i ( x) and f upper
( x)   c i  i ( x) where lower,
1 1

T   x x x2 x3 x4 x5 x6 x7 x8 
Results and conclusion

An in-house compressible RANS code under development has been used to compute the
aerodynamic characteristics of the four models. The code runs with tetrahedral meshes obtained
with TetGen and employs a cell-centered scheme. A fourth order Runge-Kutta scheme is used for
time advancing and a second order scheme for the spatial discretization. Spalart-Allmaras
turbulence model is used. Roe scheme is used as interior scheme and also for the far field
boundaries, where the fluid domain is about 20 wing spans. The solid wall boundary condition is
using a wall law [3]. Meshes are quite coarse, around 100K nodes and 600K cells. A mesh
sensitivity comparison has been done for the fourth configuration. Thus, the 5th configuration is in
fact the same as the 4th, with a refined mesh consisting of 300K nodes and 1.5M cells. There are
not significant differences. Apparently, the most promising configuration as L/D is config. 3.
However, the gain in L/D is too small for an increment in AR from 8.8 to 14 and related
complications. That is because of decreasing the Re number from 2.3e5 to 1.8e5.

Lift curve Drag polar

1.4
1.4
1.2
1.2
1
1
0.8
conf. 2 conf. 2
Cl[-]

conf. 1 0.8 conf. 1

Cl[-]

0.6 conf. 3
conf. 3 0.6
0.4 conf. 4 conf. 4
0.4 conf. 5
conf. 5
0.2
0.2
0
0
-1 2 5 8 11 14 17 2.00E-02 7.00E-02 1.20E-01 1.70E-01 2.20E-01

AoA[deg] Cd[-]

Lift/Drag curve Pitching moment

conf. 2
0
15 conf. 1
-1 2 5 8 11 14 17
-0.1
conf. 3 conf. 2
13
conf. 4 -0.2 conf. 3

11 conf. 5 -0.3 conf. 4

conf. 5
L/D[-]

CMy[-]

-0.4
9
-0.5
7 -0.6

-0.7
5
-0.8
3 -0.9
0 2 4 6 8 10 12 14 16
AoA[deg]
AoA[deg]

Fig. 4 Computed aerodynamic characteristics of the four configurations

International Conference of Aerospace Sciences “AEROSPATIAL 2010” ISSN 2067-8622
Bucharest, 20-21 October, 2010 3
 “AEROSPATIAL 2010” – Section 1

Assessing the graphs in Fig. 4, we get values for L/D at cruise, for all configurations in ascending
order as: 12, 11.6, 12.25, 11.75, 12. A simple calculation reveals the required mass for the battery
pack, Lithium-Polymer. This class of batteries, although good in aeromodelling, doesn’t offer us the
best energy density. In order to satisfy the requirements of flight – 2h endurance at 15m/s – for a
configuration with L/D=12, the pack weights at least 1.05Kg and needs a capacity of more than
12Ah (three cells – 11.1V). Thus, the possibilities are 2 packs of 6Ah or better 3 packs of 5Ah.

Lipo
No. Weight L/D Eta Velocity Power En mass Capacity
[daN] [-] [-] [m/s] [W] [Wh] [g] mAh
1 3 10 0.60 15.0 75.0 150.0 1250.0 13514
2 3 12 0.60 15.0 62.5 125.0 1041.7 11261
3 3 14 0.60 15.0 53.6 107.1 892.9 9653
4 3 16 0.60 15.0 46.9 93.8 781.3 8446
Table 1 Battery capacity for the given mission, as a function of L/D

The parameter Eta in Table 1 is the product of efficiencies of propeller and electric motor.
The 4th configuration, with elliptic wings and tail, appears to have almost the same performance as
the 2nd, which confirms that the simpler design is overall, better. The 3rd configuration, with a much
larger AR, is disappointing, increasing. The second configuration is to be chosen, having the same
performance, while maintaining simplicity - three panel ruled surface wing, easy to manufacture.

References

[1] Development of Unmanned Aerial Vehicle (UAV) for Wildlife Surveillance, Kyuho Lee, A thesis
presented to the graduate school of the University of Florida in partial fulfillment of the requirements for the
degree of Master of Science, 2004
[2] Low Reynolds Number Airfoil Design and Wind Tunnel Testing, Low Reynolds Number Aerodynamics,
Proceedings of the Conference Notr Dame Indiana, USA, 5-7 June 1989, ISBN 3-5 40-5 1884-3 Springer
Verlag Berlin Heidelberg New York
[3] Fluid Dynamics Computation with NSC2KE, An User-Guide, Release 1.0, Bijan Mohammadi, 1994,
INRIA

International Conference of Aerospace Sciences “AEROSPATIAL 2010” ISSN 2067-8622

Bucharest, 20-21 October, 2010 4