new winglet for ls6 · 2011. 5. 11. · original wing - total drag 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6...

New winglet for LS6

Winglet aerodynamic design experience

&

Corresponding RANS analysis using TAU (DLR)

Matthieu Scherrer & Stefan Melber-Wilkending (DLR)

Introduction Winglet for LS6 ? Physical aspects at stake for winglet design

Winglet aerodynamic design Winglet concept Parametrical study Final design

Back to back analysis of winglets vs plain tips with TAU-RANS (DLR) Presentation of the computations Local analysis Computed drag & lift analysis

Conclusion

Contents

Introduction

Winglet is a must on all last sailplane Benefits in competition proved over the years,

though not fully catched in theory

LS6a/b Probably the most optimised 15m with plain tips.

Wonderful handling & confort

Existing retrofits LS6c/WL factory WL (similar to LS8)

Piontowsky design for LS6a version (similar to factory LS6-18W)

Darlington design for LS8

Winglets for LS6 ?

Aerodynamic aspects to be considered in winglet design

Expected performance benefit of winglet is small Sailplane are already high AR, low induced drag ships

Even “side effects” to be considered

List of items to be considered in winglet design Effect on induced drag (equivalent AR)

& Additional wetted area

Multipoint design

Effect of winglet on wing tip transition pattern

Trimming drag & winglets

It is reported that re-laminarization due to winglet plays an important role in the benefit on the speed polar Principle sketch

Can CFD catch this phenomenon ?

Change in wing tip transition pattern

tr

Tail loading affects the sailplane wake, hence the induced drag Detectable effect on the benefit brought by winglet

Trimming drag

Wingtlet benefit DCD: -2.21% Wingtlet benefit DCD: -1.72%

Winglet design

Winglet concept

Two extrem sort of existing winglets

Big “shovel” Robustness to sideslip

Plain tip + winglet blade High efficiency winglet

It was choosen to have a high AR blade plus a large transition region

Design parameters

Height/span Retrofit limitation

Chords Reynolds vs wetted area

Sweep angles

Local loading

Dihedrals

Blade angle for wing flexion Transition radius

Twist / toes

Local loading ->3DOF to be optimised

-0.5

-0.25

0

0.25

0.5

0.75

1

1.25

1.5

6.000 6.500 7.000 7.500 8.000

-0.5

-0.25

0

0.25

0.5

0.75

1

1.25

1.5

6.000 6.500 7.000 7.500 8.000

Custom airfoil design Derived from model glider experience

Forward pressure recovery gradient for robustness to separation

Main winglet airfoil design

CL=0.6

-1

-0.75

-0.5

-0.25

0

0.25

0.5

0 0.25 0.5 0.75 1

x/c

Cp

PSU94097 (ref)

MS-Wglt

Parametric study

4DOF study Geometric : Outer wing twist, winglet twist & toe

Aerodynamic : CL (0.2, 0.6, 1.2)

Use of rapid – relatively simple tool Induced drag is computed without wake balancing (AVL)

Viscous drag is locally evaluated along span as function of Cl, based on 2D polars.

Drag figures increment computed with trimming constraint through elevator deflection, for a mid CG.

saiplanesaiplanesaiplanewinglet CLCDCLCDCLCD sailplane trimmed

wingletw/o

sailplane trimmed

wingletw D

Raw results from parametric study

Evolution of winglet benefit vs 4DOF (3 angle & CL)

CL=0.2

CL=0.6

CL=1.2

Multipoint approach

Merit figure to be optimised : weighted polar, according to flight template theory.

“Envelope Flight template”

0 0.5 1 1.5CL

FT

(CL

)

Enveloppe Flight Template

Represents a statistically relevant program of a typical cross country flight. It is used for defining a multipoint cost function :

LLt

L

LD

L

D dCCfC

CC

C

C

range C 2/32/3L

)()(

Ref paper ostiv

Parametric study with multipoint approach

Merit figure to be optimised : weighted polar, according to flight template theory.

Target wing loading approach

Local load along span Local Cl should not exceed the max lift of airfoil

Helps at selecting toe/twist combination among quasi equivalent solutions

Local lift distributions

0

0.25

0.5

0.75

1

1.25

5.5 6 6.5 7 7.5 8

Y

Cl

Local Cl, Ref LS6 (no WL), CL=0.2Local Cl, Ref LS6 (no WL), CL=0.6Local Cl, Ref LS6 (no WL), CL=1.2Local Cl, LS6WL, CL=0.2Local Cl, LS6WL, CL=0.6Local Cl, LS6WL, CL=1.2Local Cl, LS6WL final tw ist/toe, CL=0.2Local Cl, LS6WL final tw ist/toe, CL=0.6Local Cl, LS6WL final tw ist/toe, CL=1.2

Rational for selecting winglets settings

The influence of the toe/twist angle seems low within a quite wide range (for the prediction method used) Typical “flat optimum range : +/-2deg

This is not in line with reported experience in WL design

This must somehow be filtered by some other criteria Local CL available from AVL

Evidence of risk of overloading in the transition region

Practical conclusion Once the winglet exists, most of its effects on induced drag is

settled (i.e. twist optimisation for Cdi is exagerated refinement)

Winglet setting should be sized for controlling viscous drag in extreme cases : intrados laminarity at low CL & extrados corner separation at high CL.

Final winglet geometry

Twist & toe Risk mitigation vs transition region loading

Not absolute optimal for AVL, local minimum

CAD for meshing & milling purpose (Flybiwo)

Swetted/Sref=1

Non-performance sizing aspects

Wing root bending moment

Handling quality

Flutter Measurement of sailplane modes sensitivity to additional masses

planned

Analysis of winglets vs plain tips using TAU-RANS (DLR)

Computation presentation Reference wing winth plain tips

Effect of winglets

CFD for sailplane purpose

Heavy CFD is rapidly developping for aerospace industry

TAU is an unstructured RANS code from DLR, that is used daily at Airbus

Can those developments give interesting results for sailplane design ? A glider is not an airliner !

Ingredient for this test case : 3D aero, laminarity

Geometry & mesh

Half a wing was meshed (symmetrical cases) Around 7.5 E6 points in the meshes

Setting up of computations

A list of 7 points along the speed polar were computed

Computation attempted for 4 conditions (7x4=28 cases) :

Winglet & plain tip geometry

Full turbulent flow & with transition

A list of 7 points along the speed polar were computed Turbulence model : K-w-SST

Transition : Tolmienn-Schlichtting waves & laminar bubble detection (N=11.5)

Elements of computation performance : 1 case = 80h on 40 cores

case CL V (m/s)

1 0.2 56.6

2 0.6 32.7

3 0.8 28.3

4 1.2 23.1

5 1.259 22.6

6 1.308 22.2

7 1.353 21.8

Analysis of winglets vs plain tips using TAU-RANS (DLR)

Computation presentation Reference wing winth plain tips

Effect of winglets

TAU for plain wing tip

Local Cp & transition comparison Comparison with MIAReX extended lifting line

Drag polar comparison Comparison with MIAReX computation & IDAFLIEG measurement

Evaluation of aerodynamic coefficients through two methods : Wing surface integration & far-field analysis

TAU for plain wing tip

Local Cp & transition comparison y/b=0 Comparison with MIAReX extended

lifting line, same CL

Comment General shape reproduced

No laminar bubble modelling for

TAU (with transition computation)

Undersurface at TE not behaving similarly (lam bubble)

Transition prediction differs

TAU for plain wing tip

Local Cp & transition comparison y/b= 0.28 & 0.53 Comparison with MIAReX extended lifting line

TAU for plain wing tip

Local Cp & transition comparison y/b=0.75 & 0.91 Comparison with MIAReX extended lifting line

TAU for plain wing tip

Comparion of transition predictions for CL=0.6 Comparison with MIAReX computation : more fwd transition line

Same definition of Transition point ?

Compatible with comparison of drag values (next slide)

Transition prediction for original wing

TAU vs MIAReX

-0.5

0

0.5

1

1.5

0 1 2 3 4 5 6 7 8y (m)

X T

r

Xtr Miarex - Upper

Xtr Miarex - Lower

Xtr Tau - Upper

Xtr Tau - Lower

TAU for plain wing tip

Results for aerodynamic coefficients

Conclusion : Absolute value from TAU cannot be used with confidence. Back to back analysis necessary :construction of an increment of

performance

Comparison of drag polars

Original wing - Total drag

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05 0.055

CD

CL

A/C CD from IDAflieg speed polar

Cxtot - Miarex

CD tot NF Turb - Tau

CD tot NF Lam - Tau

CD tot FF Turb - Tau

CD tot FF Lam - Tau

0. Two references: IDAFLIEG. Polar MIAReX computation (lifting line)

1. Full turbulent result - More drag for the wing alone than for

sailplane as measured by IDAFLIEG. - Difference FarField/Nearfield

1

2. Results with transition - Farfield analysis shows result more

compatible with IDAFLIEG

2

Analysis of winglets vs plain tips using TAU-RANS (DLR)

Computation presentation Reference wing winth plain tips

Effect of winglets

Effect of Winglet in the flowfield

It is possible to “explore” the flowfield in the wake of the wing & winglet

Winglet & wing tip pressure field

Three station where considered to analyse local pressure features

Cut 1

Cut 2

Cut 3

Pressure line near tip for CL=0.2

-1.25

-1

-0.75

-0.5

-0.25

0

0.25

0.5

0 0.25 0.5 0.75 1

x/c

Cp

Plain tip cut 1WL Cut 1WL Cut 2WL Cut 3

Evolution of wing tip pressure field

CL=0.2, V=56.6m/s

Full turbulent solution

Pressure line near tip for CL=0.6

-1.25

-1

-0.75

-0.5

-0.25

0

0.25

0.5

0 0.25 0.5 0.75 1

x/c

Cp

Plain tip cut 1WL Cut 1WL Cut 2WL Cut 3

Winglet & wing tip pressure field

CL=0.6, V=32.7m/s

Full turbulent solution

Pressure line near tip for CL=0.8

-1.5

-1.25

-1

-0.75

-0.5

-0.25

0

0.25

0.5

0 0.25 0.5 0.75 1

x/c

Cp

Plain tip cut 1WL Cut 1WL Cut 2WL Cut 3

Winglet & wing tip pressure field

CL=0.8, V=28.3m/s

Full turbulent solution

Pressure line near tip for CL=1.2

-2.5

-2.25

-2

-1.75

-1.5

-1.25

-1

-0.75

-0.5

-0.25

0

0.25

0.5

0 0.25 0.5 0.75 1x/c

Cp

Plain tip cut 1WL Cut 1WL Cut 2WL Cut 3

Winglet & wing tip pressure field

CL=1.2, V=23.1m/s

Full turbulent solution

Winglet & transition line

Effect of winglet on transion line is represented by CFD

Delay the transition by up to 8% on pressure side

Phenomenon damped over 4% of span

Transition prediction with & w/o winglet

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

6.8 6.9 7 7.1 7.2 7.3 7.4 7.5 7.6

y (m)

X T

r

Xtr Tau - Upper

Xtr Tau - Lower

Tau WL - Upper

Tau WL - Lower

Transition prediction with & w/o winglet

0%

10%

20%

30%

40%

50%

60%

70%

6.8 6.9 7 7.1 7.2 7.3 7.4 7.5

y (m)

X T

r

Xtr Tau - Upper

Xtr Tau - Lower

Tau WL - Upper

Tau WL - Lower

Winglet benefit evaluation

Drag modification brought by winglet

NB : Reliable winglet computation with transition : only CL=0.2 & 0.9, higher CL extrapolated

Effect of winglets

Drag benefit

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

-0.0015 -0.001 -0.0005 0 0.0005

DCD

CL

DCD winglet (full turb)

DCD winglet (Transition computation)

Effect of winglets

Relative drag benefit

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

-4.00% -3.00% -2.00% -1.00% 0.00% 1.00% 2.00% 3.00%

DCD

CL

DCD winglet (full turb)

DCD winglet (Transition computation)

Winglet benefit evaluation

Pilot perspective

NB : Reliable winglet computation with transition : only CL=0.2 & 0.9, higher CL extrapolated

Effect of winglets - Pilot perspective

-1

-0.5

0

0.5

1

1.5

2

50 100 150 200 250

V (kph) @ 33kg/m²

D L

/D

Increment of L/D due to winglet(full turb)Increment of L/D due to winglet(Transition computation)

0.5

0.75

1

1.25

1.5

70 90 110 130 150 170

V (kph) @ 33kg/m²

Vz (

m/s

)

IDAFLIEG

WL with benefit (full turbulent computation)

WL with benefit (computation with transition effect)

Conclusion

Conclusion

Design of new winglet for LS6 A concept as starting point

Simple tools & pragmatic approach for angle setting

Winglet analysis Use of available method, over a quite large number of

computation cases

Understanding of local aerodynamic of winglet

Documented evaluation of performance modification

Next step : let’s fly those winglets !

Special thanks

DLR (TAU code)

Flybiwo (CAD & milling)

Have nice flights !

Questions ?

matthieu.scherrer@free.fr