Journal of Physics: Conference Series

PAPER • OPEN ACCESS Related content

- Solar Cell to Support Perpetual Flight of
Comparative Study of Wing Lift Distribution High Altitude Long Endurance UAV ITB

Analysis for High Altitude Long Endurance (HALE) - Design of High Altitude Long Endurance
UAV: Structural Analysis of Composite
Wing using Finite Element Method
Unmaned Aerial Vehicle
- Longitudinal Static Stability and wake
visualization of high altitude long
To cite this article: Faber Y. Silitonga and M. Agoes Moelyadi 2018 J. Phys.: Conf. Ser. 1005 012036 endurance aircraft developed in Bandung
institute of technology

View the article online for updates and enhancements.

This content was downloaded from IP address 106.51.231.69 on 24/08/2020 at 12:05

5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

Comparative Study of Wing Lift Distribution Analysis for

High Altitude Long Endurance (HALE) Unmaned Aerial
Vehicle
Faber Y. Silitonga1)* and M. Agoes Moelyadi1)**
1)
Department of Aeronautics and Astronautics, Faculty of Mechanical and Aerospace
Engineering, Institut Teknologi Bandung, Indonesia.

*[email protected] **[email protected]

Abstract. The development of High Altitude Long Endurance (HALE) Unmanned Aerial
Vehicle (UAV) has been emerged for both civil and military purposes. Its ability of operating
in high altitude with long endurance is important in supporting maritime
applications.Preliminary analysis of HALE UAV lift distribution of the wing presented to give
decisive consideration for its early development. Ensuring that the generated lift is enough to
compensate its own weight. Therotical approach using Pradtl’s non-linear lifting line theory
will be compared with modern numerical approach using Computational Fluid Dynamics
(CFD). Results of wing lift distribution calculated from both methods will be compared to
study the reliability of it. HALE UAV ITB has high aspect ratio wing and will be analyze at
cruise flight condition. The result indicates difference between Non-linear Lifting Line and
CFD method.

Keywords: aerodynamic, lift distribution, HALE UAV.

Nomenclatures
L : Lift
L’ : Lift distribution (per chord)
C : Chord
Cl : 2D Lift coefficient
CL : 3D Lift coefficient
ρ : Density
P : Static pressure
V : Velocity
S : Wing are
q : Dynamic pressure
α : Angle of attack
b : Span
AR : Aspect ratio
x,y : Coordinates

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

1. Introduction
Indonesia is a very large country that contains of vast amount of islands, not only a huge archipelago,
also rich of natural resources the spread from the land throughout the ocean. Large territory which
holds thousands of island, challenges to provide modern technology that able to monitor its large area
within certain amount of times. Not only being able to monitor large area with long endurance, human
resources availability needs to be taken as consideration. HALE UAV ITB (Institut Teknologi
Bandung) development is to fulfill the need of reliable autonomous vehicle in order to monitor
Indonesian country border and natural resources, that can also be controlled by remote control and
telecommand from ground station. ITB bas three years plan to develop HALE UAV that divided into
three stages. The first stage produced a 12 m wingspan prototype with four boom configuration with
four electric propulsion system for about 20000 ft altitude, as shown in figure 1. First prototype had
some problems with side wind effect that cause heavy bending moment, damaging the wing. The
second stage, which is currently ongoing project, will have a 16 m wingspan that able to operate at
20000 ft altitude. The second prototype, showed in figure 2, will be using single boom with dual
electric motor, which is obviously provides smaller weight than four boom configuration. Larger
wingspan means larger aspect ratio, which expected to be producing spanwise lift distribution more
similar with rectangular lift distribution approach due to relatively smaller induced drag produced. If
the hypothesis is satisfied thus, winglet addition can be ignored that will reducing wing bending
moment caused by side wing effect.

Spanwise lift distribution prediction can be used to predict the structural load, which is important in
designing the structural initial sizing and configuration of the wing. Prandtl’s Non-linear Lifting Line
Theory, is choosen rather than Prandtl’s Classic Lifting Line Theory due to is ability to predict the
non-linear Cl vs alpha curve that usually happened at high angle of attack thus, expected to produce
more accurate result with relatively less time consumed, to CFD simulation. On the other hand, CFD
requires longer amount of time but calculate more accurate lift distribution. The comparison between
both method will test the reliability of Prandtl’s Non-linear Lifting Line Theory ability to predict the
spanwise lift distribution of HALE UAV ITB.

Figure 1. First generation of HALE UAV-ITB prototype

2. HALE UAV Configurations

2.1. Basic Configurations

HALE UAV-ITB has high wing and single tail-boom configuration with vertical and horizontal tail as
stabilizer. The length is 4 m with 16 m wing span. Airfoil that been used for the wing is EMX 07 and
for the stabilizers are NACA 0012.

2
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

Table 1. HALE UAV-ITB Configuration

MTOW 20 kg
Wingspan 16 m
MAC 0.4 m
Taper Ratio 1
Twist Angle 0 degree
Aspect Ratio 40
S (Wing Area) 6.4 m2
Altitude 20000 ft

Figure 2. HALE UAV-ITB three view drawing

a.)
EMX 07
thickness/chord

0.1
0.05

0
0 0.2 0.4 0.6 0.8 1
-0.05
x/chord

b.)
NACA 0012
thickness/chord

0.1
0.05
0
0 0.2 0.4 0.6 0.8 1
-0.05
-0.1
x/chord

Figure 3. Airfoil: a.) EMX-07 b.) NACA 0012

2.2. Flight Configurations

For surveillance, the designed HALE UAV ITB will fly at cruise stage as follow:

3
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

Table 2. HALE UAV ITB Flight Condition

Cruise speed 60 km/hour
Angle of Attack (α) 3 degree
Altitude 20000 feet
Air density 0.73 kg/m3
Ambient Temperature 250 K
Ambient Pressure 46650 Pa

3. Prandtl Non-linear Lifting Line Theory

The following procedure are Prandtl’s Non-linear Lifting Linear Theory method by Anderson[1]. Wing
will be divided into a number of k+1 spanwise stations and the initial circulation (Γ) value, at given
angle of attack (α), for each station will be assumed as elliptical lift distribution. Then, the induced
angle of attack αi for each section will be calculated by solving the following equation:

⁄𝟐 𝒃
𝟏 (𝒅𝚪⁄𝒅𝒚) 𝒅𝒚
𝜶𝒊 (𝒚𝒏 ) = ∫ (1)
𝟒𝝅𝑽∞ −𝒃⁄ 𝒚𝒏 − 𝒚
𝟐

Using effective angle of attack (αeff) at each station can be calculated by using αi from equation (1) and
the wing condition angle of attack, by solving the following equation:

𝜶𝒆𝒇𝒇 (𝒚𝒏 ) = 𝜶 − 𝜶𝒊 (𝒚𝒏 ) (2)

Now, a new Γ distribution is obtained by the latest value of effective angle of attack, and will be
compared to the previous value of Γ. If a difference occur between the old and new value of Γ, a new
input value of Γ, generated from 𝚪𝒊𝒏𝒑𝒖𝒕 = 𝚪𝒐𝒍𝒅 + 𝑫 ∙ (𝚪𝒏𝒆𝒘 − 𝚪𝒐𝒍𝒅 ) where D is a damping
coefficient[1], to perform iteration until the latest value of Γ agree with its previous value within
specified margin of error. After that, Kutta – Joukowski theorem, shown by equation (3), can be
performed using the Γ value to generate lift at each section.

𝟏
𝑳′ (𝒚𝒏 ) = 𝝆∞ 𝑽∞ 𝜞(𝒚𝒏 ) = 𝝆 𝑽 𝟐 𝒄 (𝒄 ) (3)
𝟐 ∞ ∞ 𝒏 𝒍 𝒏

Total lift that produced by the wing will be obtained by integrating the value of 𝑳′ (𝒚𝒏 ) spanwise.

4. Computational Fluid Dynamics (CFD) Simulation

Computational Fluid Dynamics (CFD) will be used as the numerical approach to calculate the
spanwise lift distribution of the wing. In this simulation ANSYS CFX will be used as the solver of
CFD simulations. CFX solver calculation is based on Reynolds Averaged Navier Stokes (RANS)
method, which is derived from Direct Navier Stokes (DNS) method. Instead of Large Eddy Simulation
(LES) method, RANS method was choosen because it delivers enough accuracy within faster
processing rather than LES, even though it has higher accuracy but takes a lot of memory and running
time. The following equations are the governing equations used by ANSYS CFX Solver, including
continuity (4), momentum (5) and total energy (7):

4
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

Continuity eq. 𝝏𝝆
+ 𝛁 ∙ (𝝆𝑽) = 𝟎 (4)
𝝏𝒕

Momentum eq. 𝝏(𝝆𝑽)

+ 𝛁 ∙ (𝝆𝑽 × 𝑽) = −𝛁𝒑 + 𝛁 ∙ 𝝉 + 𝒔𝑴 (5)
𝝏𝒕

Stress tensor,τ, 𝟐
strain rate relation 𝝉 = 𝝁(𝛁𝑽 + (𝛁𝑽)𝑻 − 𝜹𝛁 ∙ 𝑽) (6)
𝟑

Total energy eq. 𝝏(𝝆𝒉𝒕𝒐𝒕 ) 𝝏𝒑

− + 𝛁 ∙ (𝝆𝑽𝒉𝒕𝒐𝒕 )
𝝏𝒕 𝝏𝒕 (7)
= 𝛁 ∙ (𝝀 𝛁𝑻) + 𝛁 ∙ (𝑽 ∙ 𝝉) + 𝑽 ∙ 𝒔𝑴 + 𝒔𝒈

Total enthalpy, htot 𝟏

𝒉𝒕𝒐𝒕 = 𝒉 + 𝑽𝟐 (8)
𝟐

Shear Stress Transport (SST) method will be used to evaluate the simulation turbulence viscous effect.

4.1. Simulation Model

CFD simulation method is done by the following workflow scheme:

Wing design Structured Defining Solving CFD

modelling meshing domain simulation

Comparing CFD Obtaining Integrating pressure

results with NLLT spanwise lift distribution per wing
results distribution section (L’)

Figure 4. CFD workflow scheme

The simulation will be simplified as a wing-only half spanwise thus, means actual total lift produced is
two times the lift calculated by the simulation. The simulation will be using viscid assumption, to
compare the lift distribution result of viscid assumption, by CFD, with invisced assumption,by Non-
linear Lifting Line Theory. Structured mesh blocking method by ICEM CFD will be used to the wing
simulation domain. Computational domain that being used are:
1) INLET boundary condition in front of leading edge
2) SYMMETRY boundary condition for wing root side of the domain
3) WALL boundary condition for the wing itself
4) OPENING boundary condition for the other domain.

5
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

Figure 5. Side view of generated meshes for wing domain

Figure 6. Computational domains for wing

5. Result

Wing will be divided into 10 section for extracting the CFD results (y = 0, 0.8, 1.6, 2.4, 3.2, 4.0, 4.8,
5.6, 6.4, 7.2, 7.95 m) to obtain Lift/chord (L’) for each station which will be used to obtain the
spanwise lift distribution. Chord wise static pressure distribution for each section will be extracted,
then by integrating the pressure distribution for upper and lower chord we will obtain L’ for each
section.

Here are some notable pressure distribution graph alongside airfoil chord at section y = 0, 7.2, and
7.95 m

6
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

Static Pressure Chordwise Distribution (y = 0 m)

54250
54200
Static Pressure (Pa)

54150
54100
54050
54000
53950
53900
53850
-0.2 0 0.2 0.4 0.6 0.8 1 1.2
x/chord
Lower Static Pressure Upper Static Pressure

Figure 7. Static pressure distribution chordwise at section y = 0 m

Static Pressure Chordwise Distribution (y = 7.2 m)

54200

54150
Static Pressure (Pa)

54100

54050

54000

53950

53900

53850
0 0.2 0.4 0.6 0.8 1 1.2
x/chord
Lower Static Pressure Upper Static Pressure

Figure 8. Static pressure distribution chordwise at section y = 7.2 m

7
 5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

Static Pressure Chordwise Distribution (y = 7.95 m)

54100
54080
54060
Static Pressure (Pa)

54040
54020
54000
53980
53960
53940
53920
0 0.2 0.4 0.6 0.8 1 1.2
x/chord
Lower Static Pressure Upper Static Pressure

Figure 9. Static pressure distribution chordwise at section y = 7.95 m

By integrating the pressure distribution, lift/chord (L’) for each section will be obtained.

Table 3. Lift/chord for each section

Section 1 2 3 4 5 6 7 8 9 10 11
y (m) 0 0.8 1.6 2.4 3.2 4 4.8 5.6 6.4 7.2 7.95
L' (N/m) 18.97 18.90 18.88 18.87 18.84 18.78 18.76 18.76 18.60 17.78 6.26

a.)
Half Spanwise Lift Distribution
20
18
16
14
Lift/Chord (N/m)

12
10
8
6
4
2 y (m)
0
0 2 4 6 8 10
CFD Non-Linear Lifting Line

8
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

b.) Half Spanwise Lift Distribution

20
18
16
Lift/Chord (N/m)

14
12
10
8
6
4
2
2y/b
0
0 0.2 0.4 0.6 0.8 1
Non-Linear Lifting Line CFD

Figure 10. NLLT and CFD lift distribution comparison a.) CFD as data per section b.) CFD as
curvature

By integrating the spanwise lift distribution and times it by two we can obtained the total lift
performed by whole wing.

Table 4. Total Lift produced and margin of difference by NLLT and CFD method
Total Lift
CFD Method 288.38 N
NLLT Method 292.59 N
Margin of Error 1.46 %

6. Analysis and Discussion

There is a slight difference of total lift calculated by CFD and NLLT with very little margin of error,
around 1.46% , which is still an acceptable value. The distribution of lift spanwisely, according to
Figure 10, shows very littly difference distribution besides the wing tip area. Those slight difference
are suspected to be caused by mathematically error of the integrating method that been used, which is
trapzoidal rule, in order to achive lift/chord for each section. But the trend of distribution is quite the
same and still acceptable. It is also due to the simulation that operated in low angle of attack, which is
still within the linear region of Cl vs alpha curve thus, lead to more accurate NLLT calculation. Lift
distribution differences between both method around wing tip are suspected to be caused by the
method of calculating induced drag, which is caused by vortex that created by flow leakage from
lower to upper wing. CFD calculation relatively simulate more accurate results towards real condition,
due to it’s ability to calculate the static pressure distribution then derive it into induced drag effect to
lift distribution. In other hand, NLLT calculate the circulation instead of static preasure distribution
which caused less accuracy than CFD calculation. But, the difference between both method is still

9
5th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1005 (2018)
1234567890 ‘’“” 012036 doi:10.1088/1742-6596/1005/1/012036

small for this case due to the wing configuration that has high aspect ratio and constant chord size
without any twisting.

7. Conclusion

The application of NLLT to predict wing spanwise lift distribution for preliminary analysis is
reccomended for HALE UAV, due to it’s assured accuracy, small margin of error (1.46%) against
CFD calculation results, and also requires relatively easier and shorter amount of time to perform.

Acknowledgements

This research was supported by ITB and Direktorat Riset Teknologi dan Pendidikan Tinggi
(RISTEKDIKTI).

References

[1] Anderson, John D. Fundamentals of Aerodynamics (Fourth Edition). McGraw-Hill Education,

New York. 2009. (Book)
[2] Anderson, John D. Computational Fluid Dynamics. McGraw-Hill Education, New York. 1995.
(Book)
[3] Carlos E. S. Cesnik, Patrick J. Senatore, Weihua Su, Ella M. Atkins. X-HALE: A Very Flexible
UAV for Nonlinear Aeroelastic Tests. The University of Michigan. 2010. (Paper)
[4] NACA 0012, http://airfoiltools.com/airfoil/details?airfoil=n0012-il
[5] EMX-07, http://www.aerodesign.de/profile/profile_sp.htm#emx07
[6] Rapinett, Annabel. Zephyr: A High Altitude Long Endurance Unmanned Air Vehicle..
Department of Physics University of Surrey.2009 (Paper)

10