CERTIFICA

1
Sensitivity Analysis of Aerodynamic Performance
of Airfoils Used in Small Wind Turbines
Jorge Antonio Villar Alé
Ricardo Schuh Frantz
Felipe Weigel
PUCRS University Brazil
villar_at_pucrs.br
António Manuel Gameiro Lopes
ADAI-LAETA
Coimbra University Portugal
antonio.gameiro_at_dem.uc.pt
Svend Enevoldsen
HS Wind Randers Denmark
svend_at_ecology.dk
2
Sensitivity Analysis of Aerodynamic Performance
of Airfoils Used in Small Wind Turbines
OBJETIVE
Applies a Blade Element Theory Code to predict
the performance of small wind turbine.
3
METHODOLOGY
1. Manufacturer Information's
2. Lift and Drag Data
Wind Tunnel and Tools
3. Viterna Corrigan Correction
4. 3D Stall Delay Corrections
5. BEM Code predictions
6. Results and conclusions
4
Main interests in this study was to evaluate

• (i) Lift and drag coefficient's to the same
  airfoil
• From airfoil design tools.
• From experimental data.
• (ii) The sensitivity of CL and CD to different
  thicknesses.
• (iii) Stall delay influence in the results

Airfoil 14
Airfoil 15
Airfoil 18
5
1. Manufacturer Information's
6
Wind Turbine Performance
Manufacturer Specifications
Power Curve
25 KW
Number of blade 3
Rotor diameter 13 m
Rated power 25 kW
Hub height 18 m
Power regulation Stall
Cut-in wind speed 4,0 m/s
Rated wind speed 12 m/s
Cut-out wind speed gt 25 m/s
Rated power rotation 65 rpm
7
Blade Twist and Chord Distribution
8
1. Manufacturer Information's
2. Lift and Drag Data
Wind Tunnel and Tools to obtain data
The manufacturer uses a airfoil with 15
thickness
Remarks
Airfoil 15 thickness
No Wind Tunnel Data
9
Lift an Drag Data
CL, CD
Airfoil 15 thickness
?

• Wind tunnel reference data airfoil (1998) ()
• (2) Using codes with potential-viscous models.
• (3) Using CFD codes (Navier-Stokes solvers).

Re1.6 x106
() Fuglsang P., Ioannis A. Christian, B. Wind
Tunnel Test of the RISØ-1 Airfoil (1998)
Risø-R-999(EN).
10

1. Wind tunnel reference data airfoil

() This airfoil was for use on the outer part of
blades for stall regulated wind turbines with a
Re 1 and 2 million
Airfoil A1 14 thickness
Experimental data of airfoil 14 provided by the
manufacturer
Airfoil A1 18 thickness
Experimental data of airfoil 18 provided for
Re1.6x106 from reference
() Fuglsang P., Ioannis A. Christian, B. Wind
Tunnel Test of the RISØ-1 Airfoil (1998)
Risø-R-999(EN).
11
XFLR5 code

• Use a Inviscid vorticity panel method (modeling
  potential flow)
• Use a Integral boundary layer equations (modeling
  viscous flow)
• To angle of attack and Reynolds number provide
• Pressure distribution
• Lift and drag coefficients.

Re1.6 x106

• XFLR5 code
• Analysis tool for airfoils, wings and planes
  operating at low Reynolds Numbers.
• XFoil is a code developed 2D and 3D airfoils
  analysis
• XFLR5 code use XFoil solver with a user-friendly
  interface.
• XFLR5 is exactly the same to XFoil
• Difference XFOIL - FORTRAN code
• XFLR5 C code.

12
Airfoil 15 thickness
Re1.6 x106

• Navier-Stokes equations
• 2D incompressible.
• Turbulence Models
• Standard k-? model
• SST k-? model
• Non-structured mesh

60000 elements. Domain 10 chords away from the
airfoil.
13
(No Transcript)
14
EasyCDF
XFLR5 code
Polinomial approach
15
1. Manufacturer Information's
2. Lift and Drag Data
Wind Tunnel and Tools to obtain data
3. Viterna Corrigan Correction AR
Model to extended 2D pre stall to post stall data
using Blade Aspect Ratio (AR) corrections
16
Extension 2D and AR correction
17
Extension 2D and AR correction
Viterna Corrigan Model
18
1. Manufacturer Information's
2. Lift and Drag Data
Wind Tunnel and Tools to obtain data
3. Viterna Corrigan Correction AR
3.1 BEM Performance
19
BEM Tool
WT_Perf
Airfoil data
Blade element theory (BEM)
20
(No Transcript)
21
1. Manufacturer Information's
2. Lift and Drag Data
Wind Tunnel and Tools to obtain data
3. Viterna Corrigan Correction AR
3.1 BEM Performance
4. 3D Stall Delay Corrections
22

• Snel model
• Linderburg model
• Du and Selig model
• Dumitrisco and Cardos model
• Corrigan e Schillings model
• Chaviaropoulos and Hansen model

Engineering Models
Stall Delay Corrections
Stall delay occurs on the blades of wind turbines
causing a lift aerodynamic (3D) with magnitude
greater than that wind tunnel data (2D).
23
BEM and Stall Delay Corrections
R6,5 m
r/R0,32

• Snel model
• Linderburg model
• Du and Selig model
• Dumitrisco and Cardos model
• Corrigan e Schillings model
• Chaviaropoulos and Hansen model

Engineering Models
24
BEM and Stall Delay Corrections
3D interpolation using HARP_Opt code
Du and Selig Model
25
1. Manufacturer Information's
2. Lift and Drag Data
3. Viterna Corrigan Correction AR
4. 3D Stall Delay Corrections
Stall Delay Model Superstimate Power curve
Test adopted to run BEM Code
26
Test adopted
Airfoil A1 14 thickness
Re1,6 x106
CL
CL Experimental data (manufacturer information)
?
Re1,6 x106
CD Using XFLR 5
CD ao a1a a2a2 a3a3 ...
Polynomial approximation
Viterna Corrigan correction
( )
Stall Delay model
27
Airfoil 14 thickness
Re1,6 x106
Polinomial approach
2D Wind tunnel
Viterna Corrigan
Airfoil 14 thickness
Polinomial approach
2D Wind tunnel
Viterna Corrigan
Stall Delay Model
28
2D Airfoil 14 VC
2D Airfoil 14 VC SD
29
Stall Delay Model Superstimate Power curve
30
1. Manufacturer Information's
2. Lift and Drag Data
3. Viterna Corrigan Correction AR
4. 3D Stall Delay Corrections
Test adopted to run BEM Code
Stall Delay Model Superstimate Power curve
Efficiency of the system corrections
31
(No Transcript)
32
Procedure to estimate the efficiency of the
system. Compared all BEM power coefficients with
manufacturer power coefficient. A division
between them gives average efficiency of the
system.
A 0,048
B 0,615
C 0,534
D 0,013
33
Stall Delay Model
34
(No Transcript)
35
Airfoil 14 thickness
Polinomial approach
2D Wind tunnel
Viterna Corrigan
Power efficiency
Stall Delay Model
BEM Tool
WT_Perf
36
Operations Envelope Reynolds Number
Wind Tunnel Data Re1.6x106
37
Operations Envelope Angle of Attack
38
Operations Envelope Angle of Attack
39
1. Manufacturer Information's
2. Lift and Drag Data
3. Viterna Corrigan Correction AR
4. 3D Stall Delay Corrections
5. BEM Code predictions
40
Review of the work presented
(1) CL and CD Data from different tools
(2) CL and CD Different thicknesses.
(3) Stall Delay Model Effects
(4) Efficiency of the System
41
(1) CL and CD Data from different tools

1. Wind tunnel data

Wind tunnel 2D airfoils are still a quality base
for use in BEM codes
(2) Potential-viscous - XFLR5 code
(3) Navier-Stokes - Easy CFD code.

XFLR5 perform best results than EasyCFD compared
with Exp. data.
Best option in this work Hybrid configuration
Airfoil 14 thickness
Polinomial approach
2D Wind tunnel
42
(2) CL and CD Different thicknesses.
Airfoil 14
Airfoil 15
Airfoil 18
Increased thickness shows increase in the drag
curve, especially in the area of post stall. This
is reflected in a fall in the curve of power for
speeds above the rated speed of the turbine.

• Experimental data of airfoil 15 thickness were
  not available.
• We used experimental data of other airfoils
  suitable for stall regulated WT
• Airfoil Risø E1-14 (14 thick)
• Airfoil Risø A1-18 (18 thick).

The best performance was Airfoil Risø E1-14.
43
(3) Stall Delay Model Effects

• The Du and Selig model applied in the work
  results in increased power curve.
• This result overestimates the manufacturer power
  curve.

The best power performance prediction are
obtained using 3D Stall Delay Model and
efficiency correction
44
(4) Efficiency of the System

• This work proposes a method to consider the
  global efficiency of WT
• Related BEM power curve with electric power curve
  of the manufacturer.
• In this work the method allows a good
  approximation with the manufacturer power curve.

Limitation It is necessary to have the
experimental curve of the manufacturer.
Limitation when you want a performance prediction
to new wind turbine.
45
The nature of the wind turbines is much more
complex than that which the models can interpret
and solve. To apply the BEM code should have
enough sensitivity in the input data for the
aerodynamic coefficients, the process of
extending post stall and stall delay correction.
Thus the alternatives engineering are facing a
dedication to a fine tuning of coefficients and
models to improve prediction of wind turbine.
46
Thanks for your attention !