Content uploaded by Ema Amalia

Author content

All content in this area was uploaded by Ema Amalia on Dec 13, 2018

Content may be subject to copyright.

(This paper to appear in IOP Journal of Physics: Conference Series)

Aerodynamics characteristics of glider GL-1 based on

computational fluid dynamics

E Amalia

1

,2, M A Moelyadi 2, R Julistina2, C A Putra2

2Faculty of Mechanical and Aerospace Engineering, Institute of Technology Bandung, Jl.

Ganesha 10 Bandung 40132, Indonesia

Abstract. Glider GL-1 is especially designed for thermal updraft condition of Indonesia.

Development of GL-1 is the first in Indonesia to design a glider for aero sport purpose in

cooperation with FASI. This glider needs a minimum aerodynamics efficiency of 8.333 to meet

design requirement derived from thermal updraft condition which needs rate of descent little

than 3 m/s at gliding angle of 2 degree. Optimum flight condition for maximum range

performance has been calculated to be 25 m/s at a condition of altitude between 300 – 2000 m

with Reynolds number of 1 – 1.5 million. Computational Fluid Dynamics (CFD) is employed

to do numerical analyses to predict aerodynamics characteristics of the glider. CFD by using

half-glider meshing results maximum lift coefficient of 1.326 at angle of attack of 8 degree,

and maximum aerodynamics efficiency of 19 at angle of attack 2 degree. Result of CFD by

using full-glider meshing gives maximum lift coefficient of 1.2556 at angle of attack of 10

degree and maximum aerodynamics efficiency of 16 at angle of attack of 2 degree. Both of

half-glider meshing and full-glider meshing are employing k- turbulence model. Comparison

with preliminary design result and benchmarking with similar gliders data is also done.

1. Introduction

FASI or Indonesia Aero Sport Federation is a house of aero sport activities in Indonesia. Its activities

are including aeromodelling, motorized flight, parachuting, hang gliding, micro-light flight, and

sailplaning/gliding. According to [1], recently, animo to aero sport in Indonesia, specifically gliding is

very big. This conclusion comes from interview result with practitioner of aero sport of FASI.

However, there is a limitation in glider that can fly well in thermal updraft condition of Indonesia. In

FASI, only the glider of Schweizer SGS 1-26 that serve gliding activity frequently. The SGS 1-26 is

manufactured by Schweizer Aircraft Company, United States and was designed on 1950’s. Because of

SGS 1-26 is a relatively old, it has low performance. That is why the gliding achievement of Indonesia

remains in national level. A newer glider available in FASI is ASW 20 which manufactured by

Schleicher Centrair, Germany. The ASW 20 is rarely flown in FASI because its performance is not

good in thermal updraft condition of Indonesia which is narrow and weak. It was said that most of

thermal updraft in Indonesia has 0.5– 3 m/s vertical speed, with about 200 – 300 m in diameter and

about 5000 – 6000 feet maximum height. With situation explained above, there is a strong need to

design and manufacture a national glider of Indonesia for aero sport activity in Indonesia to achieve

better achievement in aerosport competition.

1

Corresponding author: ema@ftmd.itb.ac.id

Effort to design a national glider is conducted and still under progress. The national glider of

Indonesia named GL-1 is a glider that having configuration like a conventional low speed aircraft. It

has a high wing and a T-tail and having only one payload who is the pilot. With the range of pilot’s

weight of 70 to 110 kg, it will make a movement of centre of gravity of 9 centimetres. Wing of GL-1

is having high aspect ratio and taper in near-tip wing portion. Figure 1 shows three-view drawing of

glider GL-1 and table 1 gives data of glider GL-1. “Gajah Layang” GL-01 is previous name of glider

GL-1. In this paper, the name of glider GL-1 will be used. More information about glider GL-1 is

available in reference [1] and following website: https://glidernasionalgl1.wordpress.com/.

From characteristic of Indonesia’s thermal updraft condition, design requirement of a national

glider GL-1 is derived as can be found in detailed in reference [1] . It should have maximum rate of

descent of 3 m/s and maximum turning radius of 150 m. This requirement leads to a minimum

aerodynamics efficiency of 8.333 for a condition of maximum rate of descent of 3 m/s for non

optimum flight as calculated with method from reference [2]. Moreover, in reference [1], [3], and [4],

some preliminary calculation of performance of glider GL-1 has been done for optimum flight derived

from condition of thermal updraft of Indonesia as mentioned above. The optimum flight condition is

for maximum range with maximum aerodynamics efficiency (CL/CD) and for maximum endurance

with minimum rate of descent. From national aerosport competition rule, for maximum range flight

condition, we take the rule for short flight which is the release altitude is 1000-2000 ft with a condition

of ready to landing when in an altitude of 500 ft. So, we take a release height for GL-1 to fly is within

1500 ft and using gliding symmetric flight method in reference [2]. As for maximum endurance flight

condition, we take the rule for endurance with a release height for GL-1 to fly is within 1500 ft with

using thermal updraft as much as possible. So, for maximum endurance flight condition, we use cross

country fligth method in reference [2]. In calculating endurance, we use an average value of thermal

updraft vertical speed of 1.75 m/s. Some performance prediction of GL-1 for optimum flight condition

of maximum range and maximum endurance are as listed in table 2. In this paper, we will only

concern with flight condition of maximum range to be analyzed by CFD because there is still a big

difference between result of reference [1] and reference [3] and [4] as could be seen in table 2.

The preliminary aerodynamics characteristic as appeared in table 2 column 2 which based on

reference [1] has been calculated by using DATCOM that mainly based on empirical data. Because

DATCOM has a limitation of making centered drag-polar only, an effort to construct uncentered drag

polar has been carried out as presented in reference [3] and [4] and used to produced performance

prediction in table 2 column 3. DATCOM is still used in preliminary design phase as a quick

estimation although the result is different with more accurate method as can be seen in reference [5]

and [6]. In reference [5], it gives pessimistic result compared to other method and shifting in CL curve.

While in reference [6], it predict different gradient of CL curve. However, the result of DATCOM is

still worthwhile to include because for evaluation of drag polar it gives reasonable agreement with

other methods as could be seen in [5] and [6].

Because now glider GL-1 has reached detail design phase, more thorough aerodynamic

characteristics evaluation is needed. Main objective of study presented in this paper is to use CFD

(Computational Fluid Dynamics) as a tool to predict aerodynamic characteristic of the glider,

especially to reach maximum flight range with a flight velocity of 25 m/s which closed enough to

prediction of reference [1] as in table 2 which is 24.3 m/s. We recalculate again the value of velocity

by using slightly different flight condition in this detail design phase. In doing CFD simulation in this

study, we use two different meshing configuration, half-glider and full-glider meshing with the same

k- turbulence model. Then, other objectives is to compare result of current study by CFD with result

of preliminary design done in reference [1], [3] , reference [4] and reference [7] and doing

benchmarking by comparing with data of existing gliders.

Figure 1. Three view drawing of glider GL-1 [1]

Table 1. Data of Glider GL-1 [1]

Wing

Vertical Tail

Horizontal Tail

Area (m2)

12

Area (m2)

0.827

Area (m2)

1.357

AR

17

AR

2

AR

5

b (m)

14.283

b (m)

1.286

b (m)

2.605

Root Chord1 (m)

0.933

Arm (m)

4.302

Arm (m)

4.315

Root Chord2 (m)

0.933

Taper Ratio

0.5

Taper Ratio

0.7

Tip Chord (m)

0.4665

Croot (m)

0.857

Croot (m)

0.613

MAC (m)

0.863

Ctip (m)

0.428

Ctip (m)

0.429

Taper Ratio

1

MAC (m)

0.6670

MAC (m)

0.526

Taper Ratio Outer

0.5

LE_swept (0)

6.4

LE_swept (0)

2.7

Leading Edge

Swept

(Outer Wing)

3.75

Dihedral(degree)

3

Twist (degree)

-3

Table 2. Performance Prediction of GL-1 from Preliminary Design in Brief

With Data of

Reference [1]

With data of Reference

[3] and [4]

In Maximum Range

Condition:

(CL/CD)max

24

30

CL

1.18

0.8

CD

0.048

0.027

Gliding angle (degree)

3.0

1.8

V (m/s)

24.3

22

t (minute)

7.5

10.4

Maximum Range (km)

11.0

13.7

In Maximum Endurance

Condition:

(RD)min (m/s)

0.76

0.66

Gliding angle (degree)

2.1

2.15

V (m/s)

18.5

17.5

Vaverage (m/s)

12.9

12.7

Range (km)

11.1

12.1

Maximum Endurance (minute)

14.4

15.9

2. Theory

In this study, derived from flight condition of glider GL-1 at maximum flight range as appeared in

reference [1], [3], and [4] and also discussed in section 1, we deal with a flow with Reynolds number

between 1 million to 1.5 million and velocity of 25 m/s. It falls into category of high Reynolds number

where turbulence occurs. In this kind of flow, the inertia forces in the fluid become significant

compared to viscous forces. This study is an applied CFD one, so that we use treatment of such kind

flow in an existing code, ANSYS CFX. For a flow with turbulence, the code provides evaluation

method by using turbulence models which consists of statistical turbulence model, large eddy

simulation, and detached eddy simulation. In this study, we choose statistical turbulence model which

is k- turbulence model. It is a “two equations turbulence model” which has advantage of good

compromise between accuracy and numerical effort. In this type of turbulence models, velocity and

length scale are treated by using separate transport equations (hence the term “two-equation”). In this

section we briefly cover about theory that underlying k- turbulence models which taken from

reference [8] and [9].

Statistical turbulence model is based on principle of modifying original unsteady Navier-Stokes

equation into average and fluctuating quantities to produce RANS (Reynolds Average Navier-Stokes)

equation. It consists of mean flow quantity only, and modelling turbulence effect without needs for

resolution of turbulence fluctuation. Statistical averaging procedure is employed to get RANS

equation. However, this averaging process introduces additional unknown terms containing products

of fluctuating quantities, which acts like additional stresses in fluid. This stress is called “Reynolds” or

“turbulent” stress, which difficult to determine and become a new unknown. Reynolds stress should be

modelled by additional equation with known quantities so that the equation could reach “closure”. The

equations used to close the system of equation determine the type of turbulence models.

RANS equation is as appeared in equation (1) and (2). Here is molecular stress and

is the

Reynolds stresses. For k- turbulence models that used in this study, we introduces two new variables

into the system of equation. Based on the eddy viscosity principle, the continuity equation is still the

same with equation (1), but the momentum equation becomes equation (3).

(1)

(2)

(3)

Where SM is the sum of body forces, is the effective viscosity by including trubulence defined by

equation (4), and p’ is the modified pressure which the definition is shown in equation (5). Last term

of equation (5) is neglected in ANSYS-CFX.

(4)

(5)

Parameters employed to construct two equations used for the system to reach “closure” are discussed

briefly in sub-section 2.1 for k- turbulence models used in this study.

2.1. k-

turbulence model

For reaching closure, k- turbulence model use the gradient diffusion hypothesis to relate the Reynolds

stresses to the mean velocity gradients and the turbulent viscosity. The turbulent viscosity is modelled

as the product of a turbulent velocity and turbulent lengths scale. The turbulent velocity scale is

computed from the turbulent kinetic energy, which is provided from the solution of its transport

equation. The turbulent length scale is estimated from two properties of the turbulence field, usually

the turbulent kinetic energy and its dissipation rate. The dissipation rate of the turbulent kinetic energy

is provided from the solution of its transport equation. The k- turbulence model relates turbulent

viscosity to turbulent kinetic energy and dissipation rate with equation (6). In this study, we use

default k- turbulence model provided by ANSYS-CFX solver.

(6)

3. Methodology

This study is a continuation of CFD study of wing of glider GL-1 which part of its results is published

in reference [7]. Wing of glider GL-1 has lift coefficient of 0.7498 with total lift of 3,258.34 N at zero

angle of attack. Methodology of this study is as shown in figure 2 and will be explained in this section.

Step 1 in figure 2 is to have geometry of glider GL-1 in IGES from CATIA or SolidWork. Figure 3

shows the geometry we use in this study. Step 2 in figure 2 is geometry repairing for meshing.

Original geometry file often has some gaps that need to fix so that the geometry is smooth and conti-

Figure 2 Methodology used in this study

1. Aircraft full configuration’s

Geometry Input

2. Geometry Repairing for meshing

3. Mesh building by using ANSYS ICEM

and employing half glider with symmetry

4. Numerical Calculation by using ANSYS CFX

and employing k- turbulence model

5. Analysis of calculation result by using

ANSYS CFD-Post and comparing with result

of preliminary design of reference [1],

reference [3], reference [4], and reference [7]

There is

discrepancy

6. Mesh building by using ANSYS ICEM

and employing full glider meshing

7. Numerical Calculation by using ANSYS CFX

and employing k-

turbulence model for full glider

meshing

8. Analysis of calculation result by using ANSYS

CFD-Post and comparing with result of

DATCOM from conceptual design, result of

preliminary design of reference [1], [3], [4],and

[7],result of half-glider.

9. Benchmarking with data of existing

gliders

Figure 3 Geometry of Glider GL-1 as input

Figure 4 Computational domain for meshing in Step 3 of figure 2

Figure 5 Unstructured grid on the surface of glider GL-1 by using half-glider model

nuous to apply meshing on it. As stated in step 3 on figure 2, meshing process in this step is using half

glider by applying plane of symmetry, so that computational domain is such as in figure 4 and

unstructured grid on the surface of the glider is as illustrated in figure 5.Total mesh by using half-

glider meshing is about 2 million elements.

Step 4 in figure 2 is employing ANSYS-CFX as solver for CFD analysis. Governing equation is

RANS with k- turbulence model as explained in section 2. Free stream velocity is 25 m/s. Boundary

condition is inlet at front, outlet at back, symmetry at left, and pressure far-field at right, top, and

bottom by refer to figure 4. A convergence criterion to achieve is 10-4. Numerical simulation is carried

out at angle of attack -6 to 12 degree with an increment of 2 degree. Results of step 4 and 5 of figure 2

will be discussed in section 4.

Because of there is a discrepancy between result of CFD by using half-glider meshing and result

from preliminary design as in reference [1], [3], and [4] and also there is a plan to do CFD simulation

with side slip and rudder deflection as suggested in reference [10], a CFD simulation by using full-

glider model of meshing is designed. Step 6 of figure 2 is mesh building by using full-glider model

constructed from geometry drawing as illustrated in figure 3. Figure 6 shows computational domain

used in step 6 and figure 7 gives an illustration of unstructured grid result. The unstructured grid with

full glider is consisting about 6 million elements.

Figure 6 Computational domain used in step 6 of figure 2

Step 7 in figure 2 is to carry out numerical simulation for full glider mesh as shown in figure 7 for

angle of attacks of 0, 2, 4, 8, 10, and 12 degree. We employ k- turbulence model. We use the same

velocity of 25 m/s. In step 8 of figure 2, the results will be compared with preliminary design result

and half-glider result. In step 9, result of current study will be compared with CFD result of existing

gliders or benchmarking. Results of step 7, 8, and 9 of figure 2 are discussed in section 4.

Figure 7 Unstructured grid used in step 6 of figure 2

4. Result and Discussion

In this section we present results of this study or step 4, 5, 7, 8, and 9 in figure 2. First, from step 4 and

5, we get result of numerical simulation compared with preliminary design result computed by

DATCOM of reference [1] and [7] as shown in figure 8 for lift coefficient. Then we compare result of

half-glider simulation with preliminary design of reference [1], [3], and [4] for drag polar and

aerodynamics efficiency as shown in figure 9 and 10. Because reference [3] and [4] is a reversed

engineering result, we cannot get data of CL versus angle of attack, , so we do not compare it in

figure 8.

From figure 8, we see that result of CFD with half glider meshing is having a good agreement with

result of CFD for glider wing from reference [7] which is the value of CL at 0 degree of angle of attack

is below the one of CFD for wing of the glider. This judgement is true for an aircraft with

conventional configuration as the glider GL-1 as shown in figure 1 and 3 according to reference [11].

However, for higher angle of attack than 2 degrees, CL from half-glider meshing is not following this

rule of thumb as can be seen in figure 8. Moreover, the result of CL from CFD by using half-glider

meshing is pessimistic compared to the DATCOM result of reference [1]. Result in reference [5] is

similar; there is a shifting in CL curve. In this reference, results of CL from MSES and XFOIL are

optimistic compared to DATCOM. This discrepancy could be from the fact that DATCOM is using

empirical method based on many aircraft database, and DATCOM does not include database for low

speed aircraft yet as indicated in reference [6].

In figure 9, we compare drag polar of current study of CFD by using half-glider meshing with drag

polar of GL-1 from DATCOM of reference [1], drag polar that is reversed-engineered from flight

manual data as presented in reference [3] and [4], and result of CFD wing from reference [7]. We can

see in figure 9 that result of DATCOM of reference [1] and result of current study by using half-glider

meshing is pessimistic, resulting in higher CD for the same CL if we compared to the result of reversed-

engineering of reference [3] and [4]. We can also see that part of result of DATCOM is close with

result of CFD wing of reference [7]. A similar partially good agreement in drag polar with DATCOM

result is also found in reference [5] and [6], which make DATCOM result is still worthwhile to use.

Furthermore, because the result of reference [3] and [4] based on actual gliders, we more believe the

result of CD of the reversed-engineering. Therefore, we can conclude that half-glider meshing results a

partially good prediction of CL and need further improvement in prediction of CD.

Figure 8 Lift coefficient comparison of half-glider meshing with preliminary design result

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

-10 -5 0 5 10 15

CL

Angle of Attack (Degree)

Lift Coefficient Comparison

CFD Half Glider DATCOM Result - Reference [1]

CFD Wing - Reference [7]

Figure 9 Drag polar comparison of half-glider meshing with preliminary design result

Figure 10 Aerodynamics efficiency comparison of half-glider meshing with preliminary design result

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

CL

CD

Drag Polar Comparison

CFD Half Glider DATCOM Result - Reference [1]

CFD Wing - Reference [7] Reference [3] and [4]

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80

CL/CD

CL

Aerodynamics Efficiency Comparison

CFD Half Glider DATCOM Result - Reference [1]

CFD Wing - Reference [7] Reference [3] and [4]

As for aerodynamics efficiency comparison, it is shown in figure 10. We can see that because of

prediction of CD that still need an improvement as discussed in previous paragraph, prediction of

aerodynamics efficiency by employing half-glider meshing of current study gives a lower value of 19

compared to preliminary design results of reference [1], [3], [4], and [7] which is 24 or 30. However, it

is still bigger than minimum aerodynamics efficiency of design requirement which is 8.333 for the

range of AOA 0 to 5 degree (which corresponds to CL of 0.7 to 1.2) where the GL-1 will fly.

From result of CFD by using half-glider meshing, we see a need to improve CFD model in order to

get a better prediction of CL and especially CD. We get a partially good prediction for CL with

judgement from reference [11]. Besides that, we notice that in reference [10] and [12], CFD analyses

of gliders are carried out with full-glider meshing because aerodynamics characteristic prediction will

be conducted for non-symmetric condition involving control surface such as rudder in further study.

Therefore, we decided to use full-glider meshing too to continue this study as illustrated in step 6 to 9

in figure 2 and discussed in section 3. We employ unstructured meshing as illustrated in figure 7.

First we present the result of step 8 of figure 2 which is a comparison of full-glider meshing with

preliminary design result of reference [1], [3], [4], result of CFD wing which partly presented in

reference [7], and half-glider meshing result as could be seen in figure 11 for lift coefficient. We can

briefly say that we do comparison of current study of full-glider meshing with preliminary design

result in step 8. In figure 11, we can see that prediction of CL from full-glider meshing is better than

from half-glider meshing, which is CL of aircraft is lower than CL of its wing. Although pessimistic

compared to DATCOM result, because of DATCOM has not included database of low-speed aircraft

as indicated in reference [6], we more believe the result of CFD by using full-glider meshing. Then,

the problem of predicting CL has been solved by using full-glider meshing.

Figure 11 Lift coefficient comparison of full-glider meshing with preliminary design result

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

-10 -5 0 5 10 15

CL

Angle of Attack (Degree)

Lift Coefficient Comparison

CFD Full Glider DATCOM Result

CFD Wing - Reference [7] CFD Half Glider

Figure 12 Drag polar comparison of full-glider meshing with preliminary design result

Figure 13 Aerodynamics efficiency comparison of full-glider meshing with preliminary design result

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

0.00 0.05 0.10 0.15 0.20

CL

CD

Drag Polar Comparison

CFD Full Glider DATCOM Result

CFD Wing Reference [7] CFD Half Glider

Reference [3] and [4]

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80

CL/CD

CL

Aerodynamics Efficiency Comparison

CFD Full Glider DATCOM Result

CFD Wing Reference [7] CFD Half Glider

Reference [3] and [4]

In figure 12, we add comparison with reference [3] and [4]. We cannot do comparison of variation

of lift coefficient with angle of attack in figure 11, because reference [3] and [4] do not produce this

data. From figure 12, we can see that result of full-glider meshing gives similar result with half-glider

meshing (see figure 9), which is it predicts higher CD for the same CL. To correct this, more grid close

to the surface of glider GL-1 is needed, so that the total grid will be more than 9 million. This effort is

still in progress because of problem in computer capability that available in this study. We are

improving the computer capability so that we hope in a near future, we will improve prediction of CD.

While for aerodynamics efficiency, because of lift coefficient and drag polar prediction such as

discussed previously, shows lower aerodynamic efficiency of 16 compared to half-glider meshing (19)

and preliminary design results (24 or 30) as shown in figure 13. However, it has fulfilled minimum

aerodynamic efficiency of design requirement of 8.33 for angle of attack 0 to 5 degree (which

correspond to CL of 0.65 to 1.05) where glider GL-1 will fly. If we are succeed to improve prediction

of CD by improving mesh near to the surface of glider GL-1 as described in previous paragraph, we

will get a better aerodynamics efficiency.

From step 8 in figure 2, we can conclude that full-glider meshing gives good result in predicting CL

compared to half-glider meshing and still need further improvement in predicting CD. To validate this

results, we do step 9 which is benchmarking with data of other existing gliders. Here we will use CFD

result of glider V5-Rondone from reference [12], real data of glider PW-5 Smyk from reference [13],

and flight test result of glider GROB G-103 from reference [14]. Figure 14 gives an illustration of

configuration of these gliders. We can see that all have conventional configuration with mid-wing and

T-tail except PW-5 Smyk that not having T-tail. Table 3 gives a list of some parameters emphasizing

that we do benchmarking of glider GL-1 with existing gliders in the same class.

Figure 14 Illustration of configuration of gliders for benchmarking

Table 3 Some parameters of gliders for benchmarking purpose

Glider Name

Wing Span

(m)

Wing Area

(m2)

Aspect Ratio

Glider

Length (m)

PW-5 Smyk

13.45

10.16

17.8

6.22

V5-Rondone

20

11.973

33.41

7.88

GROB 103 Twin II

17.5

17.8

17.1

8.18

GL-1

14.28

12

17

6.79

Figure 15, 16, and 17 shows benchmarking result for lift coefficient, drag polar, and aerodynamics

efficiency respectively. In figure 15 for lift coefficient, we cannot include data of PW-5 Smyk because

the data provided in reference [13] is the gradient of lift curve only. From figure 15, we could see that

result of half-glider meshing is optimistic compared to V5-Rondone and GROB 103 Twin II. It has

been discussed previously that it has been corrected by doing CFD simulation by employing full-glider

meshing. Indeed that in figure 15, result of CFD full-glider meshing agrees well with result for V5-

Rondone and GROB 103 Twin II. So, we can validate that result of CFD by employing full-glider

meshing is good in predicting CL.

Figure 15 Lift coefficient comparison for benchmarking

Figure 16 gives benchmarking result of drag polar. Here, we have also result of PW-5 Smyk from

reference [13]. From this figure we validate the fact that current study with half-glider meshing and

full-glider meshing is still need further improvement in prediction of CD. We can see, that existing

gliders in the same class is having minimum CD of 0.015 to 0.02, while half-glider meshing gives

minimum CD of 0.033 and full-glider meshing gives minimum CD of 0.043. The result of current study

is still about twice of CD value of existing gliders.

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

-8 -6 -4 -2 0 2 4 6 8 10 12 14

CL

Angle of Attack (Degree)

Lift Coefficient Comparison for Benchmarking

CFD Full Glider CFD Half Glider

V5-Rondone Reference [12] GROB 103 Twin II Reference [14]

Figure 16 Drag polar comparison for benchmarking

Figure 17 Aerodynamics efficiency comparison for benchmarking

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

CL

CD

Drag Polar Comparison for Benchmarking

CFD Full Glider CFD Half Glider

V5-Rondone Reference [12] PW-5 Smyk Reference [13]

GROB 103 Twin Reference [14]

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

CL/CD

CL

Aerodynamics Efficiency Comparison for

Benchmarking

CFD Full Glider CFD Half Glider

V5-Rondone Reference [12] PW-5 Smyk Reference [13]

GROB 103 Twin II Reference [14]

Finally yet importantly, figure 17 gives benchmarking result of aerodynamics efficiency. We can

see that maximum aerodynamics efficiency for existing gliders is in the value between 25 to 38, while

half-glider meshing gives value of 19 and full-glider meshing gives value of 16. This value is a direct

effect of prediction of CD that is still need improvement as discussed above, because the value of CL is

already in the same range with existing gliders.

5. Conclusion

Current study of CFD by employing half-glider meshing gives partially good result for CL prediction

below angle of attack of 4 degree and gives CD value about twice when compared with preliminary

design result. Half-glider meshing gives maximum aerodynamic efficiency of 19 that still below the

target of 24 or 30 from preliminary design, but above minimum aerodynamic efficiency from thermal

updraft condition (8.333). Then, the study is continued with CFD by employing full-glider meshing,

and gives improvement in prediction of CL compared with preliminary design result. However, CFD

with full-glider meshing has not improved the prediction of CD. Full-glider meshing gives maximum

aerodynamics efficiency of 16 that still below the target of 24 or 30 from preliminary design, but

above minimum aerodynamic efficiency from thermal updraft condition (8.333). This conclusion is

validated by benchmarking with existing gliders, which shows that the value of CL from current study

is in the range of the same class gliders but CD value is about twice of the value of the same class

gliders.

6. References

[1] Pratama H A 2015 Conceptual Design of Indonesian National Glider GL-01 (Bandung: Institute

of Technology Bandung).

[2] Ruijgrok, G J J 1990 Elements of Airplane Performance (Delft: Delftse Universitaire Pers)

p 306-329.

[3] Kurniasari N A D 2016 Estimation of Aerodynamics Parameter of National Glider GL-01 By

Reversed Engineering (In Bahasa Indonesia) (Bandung: Institute of Technology Bandung).

[4] Kurniasari N A D, Nurhakim M L I, Arifianto O and Mulyanto T 2016 Glider Uncentered Drag

Polar Estimation Based on Flight Manual Data Advance in Aerospace Science and Technolo-

gy in Indonesia vol 1 p 118-125.

[5] Olson E D 2015 Semi-Empirical Prediction of Aircraft Low-Speed Aerodynamic

Characteristics AIAA Science and Technology Forum (SciTech 2015)

https://ntrs.nasa.gov/search.jsp?R=20150006019 p 10.

[6] Rithice R W 2013 Comparative Evaluation of Computational Techniques for Estimating UAV

Aerodynamic, Stability and Control Derivatives (Minnesota: University of Minnesota)

https://conservancy.umn.edu/bitstream/handle/11299/157177/Ritchie_umn_0130M_13734.pdf

p 40-43.

[7] Putra C A, Julistina R, Moelyadi M A 2016 Comparative Study between Schrenk and CFD

Analysis for Predicting Lift Distribution along Wing Span of Glider Aircraft Advance in

Aerospace Science and Technology in Indonesia vol 1 p 108-117.

[8] ANSYS Inc. 2009 Ansys CFX-Solver Theory Guide (Canonsburg PA: ResearchGate) Release

12.1 p 53-61.

[9] Tenneke H and Lumley J L 1972 A First Course in Turbulence (Massachusets USA: MIT Press).

[10] Boermans L M M 2006 Research on Sailplane Aerodynamics at Delft University of Technology.

Recent and Present Developments (Delft: TU Delft) p 9-17.

[11] McCormick B W 1994 Aerodynamics, Aeronautics, and Flight Mechanics Second Edition (New

York: John Wiley & Sons) p 56-148.

[12] Guggiari T 2016 Aerodynamics of a Glider: Development of Computational Tools and Applica-

tion Studies (Milano: Politecnico di Milano) p 10-55.

[13] Rogowski K and Maronski R 2011 Optimization of Glider’s Trajectory for Given Thermal Con-

ditions The Archive of Mechanical Engineering Vol.LVIII Number1 p 18.

[14] Hendarko 1997 Flight Test to Determine Gliding Flight Performance of Glider GROB G-103

“Twin II” (in Bahasa Indonesia) (Bandung: Institute of Technology Bandung).

Acknowledgments

This study is funded by P3MI research funding of Institute of Technology Bandung. We wish to thank

to Dr. Taufiq Mulyanto and team of design group for providing data of glider GL-1.