HAWC2 Course: Lesson 3: Aerodynamic Modeling and Implementation in Hawc2

HAWC2 course
Lesson 3: Aerodynamic modeling and implementation in

HAWC2
Wind
description
Aerodynamic Aerodynamic drag forces

blade forces on tower and nacelle
Structural
model
Aeroelastic
response
2 DTU,DTU,
Risø Department
Technical of Wind Energy
University of Denmark Presentation name 17/04/2008
Overview of aerodynamic force calculation
Induction Airfoil section

modeling aerodynamics
Blade
dynamics
Aerodynamic Tower
blade forces influence
Wind
description
Main
Blade Blade Airfoil
user centerline def. planform data
input
- in htc_file - in ae_file - in pc_file
3 DTU,DTU,
Risø Department
Airfoil section aerodynamics
L
M
D
4 DTU,DTU,
Risø Department
Stall types
5 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

Lift and drag – example
L  1 2  W 2c CL ( ) D  1 2  W 2c CD ( )
Profile coefficients: how to retrieve them
• Where to get the profile coefficients from:
– 2D simulations (panel codes, CFD)
– 2D wind tunnel tests
• Limited angle of attack range:
– Extend airfoil data over a larger angle of attack

range
• 2D simulation/measurements:
– Apply 3D corrections: stall delay effects

7 Hawc2 12-14
Profile coefficients: 3D corrections
Some useful references:

– B. Montgomerie, Methods for root effects, tip effects and extending the angle of attack range
to +-180 deg, with application to aerodynamics for blades on wind turbines and propellers.
FOI - Swedish Defence Research Agency, 2004, URL: http://www2.foi.se/rapp/foir1305.pdf
– C. Lindenburg, Investigation into Rotor Blade Aerodynamics. 2003, URL:
www.ecn.nl/publications/PdfFetch.aspx?nr=ECN-C--03-025
– S.-P. Breton, F. N. Coton, and G. Moe, “A study on rotational effects and different stall delay
models using a prescribed wake vortex scheme and NREL phase VI experiment data,” Wind
Energy, vol. 11, no. 5, pp. 459-482, 2008, doi:10.1002/we.269

8 Hawc2 12-14
Life is dynamic…
2D aerodynamic model should account for dynamics (unsteady forces) in:
• Attached flow :
Memory effects of vorticity
shed into the wake.
• Flow separation (dynamic

stall):
Dynamics of the boundary
layer, delay in flow
separation.
• Non-circulatory terms:
added mass, acceleration
terms.

9 Hawc2 12-14
Results from CFD: Dynamic stall

Dynamic stall model (1)
• Lift as weighted sum:
• Clatt and Clfs computed from steady

input data.
• Dynamics from weight, separation

function f

1 Hawc2 12-14
Dynamic effects on both Cl, Cd and Cm
Please use the MHH stall

model as default since
the Øye model only
contains loops on Cl
12 DTU,DTU,
Risø Department
Induction modelling
BEM
Actuator disc
CFD
Actuator disc
Increasing Fixed wake

vortex codes Airfoil data
model
complexity
Free wake
vortex codes
CFD
Actuator line
Full
3D CFD
13 DTU,DTU,
Risø Department
Actuator disc Model of ideal turbine
- Infinitely many blades
The reaction of the blade forces applied

on the flow as volume or body forces
BEM equations (relation between

induction and load) derived for a
constant loaded actuator disc
14 DTU,DTU,
Risø Department
The energy conversion – Bernoulli
T  pA
Bernoulli
1
2 V 2  p  H
Applied before and after rotor disc
p0  12 V02  p  12 u 2
p  p  12 u 2  p0  12 u12

 p  12  V02  u12 
15 DTU,DTU,
Risø Department
The energy conversion – 1D momentum
Change in momentum
 
T m
 V  T  V0 ACVV0  u1 A1u1  V0 ( ACV  A1 )V0  m
 sideV0
Conservation of mass and insert in momentum equation
ACVV0  A1u1   ( ACV  A1 )V0  m side  side  A1 (V0  u1 )

m
  uA  u1 A1
m yields T m
 (V0  u1 )
Combining momentum and Bernoulli
u  12 (V0  u1 )
16 DTU,DTU,
Risø Department
1D momentum continued
Introducing the induction factor a
u  (1  a)V0 u1  (1  2a)V0
Extracted power for an ideal rotor is difference in power from inlet to outlet
P  12 m 
 V02  u12 
P 2 AV03a(1  a) 2
Cp    4 a (1  a ) 2
2 V0 A
1 3
Pavail
Similar for the torque
T  uA(V0  u1 )
T 2 AV02 a(1  a)
CT    4a(1  a)
Tavail 1
2 V 0
2
A
17 DTU,DTU,
Risø Department
The relation between thrust and axial induction
u  1  a V0
CT  4a1  a 
T
CT 
1 V 2 A
2 0
19 DTU,DTU,
Risø Department
Blade Element Momentum theory
Rotor plane is discretized into

independent concentric annular elements
and azimuthal variation of loading
on which the 1D assumption is applied
Relative velocity

Induction model in HAWC2
A non-rotating grid is defined in which induced velocities are

calculated. Enables modelling of azimuthal variation of induction
e.g. due to shear in inflow or due to turbulence
21 DTU,DTU,
Risø Department
BEM in HAWC2 formulated locally in grid.
Axial thrust at given

radial station on the
blade
The local thrust

coefficient is found
uinduc,axial,local
V is the magnitude of the local free wind speed. a
V ,local
22 DTU,DTU,
Risø Department
BEM in HAWC2 continued
a  k3CT 3  k2CT 2  k1CT  k0
k(3)=0.0892074
k(2)=0.0544955
k(1)=0.251163
k(0)=-0.0017077
23 DTU,DTU,
Risø Department
Tangential wake rotation
Angular momentum on annular ring
dQ   (2 r dr )(1  a)V (2a' )
Tangential blade forces on annular ring
dQ  12 W 2Cx ( )cN B r dr
Above equations solved for the
tangential induction factor
W 2C x ( )cN B
a' 
8r 2 (1  a)V 
24 DTU,DTU,
Risø Department
Sub models to the induction
modeling
 tip correction model to account for a finite

number of blades
 a yaw model to account for skew inflow
 a dynamic induction model to account for

time lag in update of wake due to load changes
25 DTU,DTU,
Risø Department
Tip loss correction model in HAWC2
The Prandtl tip correction

– take into account finite
number of blades
26 DTU,DTU,
Risø Department
University of Denmark Illustrations from deVries,
Presentation 197917/04/2008
name
Tip loss correction model in HAWC2
• The tiploss model used in HAWC2 is based on the modified expression by
Wilson and Lissaman
2   N (1  r / R)  
F cos 1  exp   B  
   2 (r / R) sin   
• This factor is inserted in the induction calculation
27 DTU,DTU,
Risø Department
Yaw correction model in HAWC2
Based on model by Glauert
28 DTU, Department
Risø DTU, ofUniversity
Technical Wind Energy
of Denmark Presentation name 17/04/2008
Azimuthal impact on yawed inflow
 Is the azimuth angle
In HAWC2 the Coleman method is used

with a constant of 0.4 instead of 0.5
29 DTU,DTU,
Risø Department
Yawed inflow
Modification of induction
acor  K a a
30 DTU,DTU,
Risø Department
Comparison between HAWC2 and CFD
actuator disc computations
30deg
60deg
31 DTU,DTU,
Risø Department
Dynamic inflow model in HAWC2
It takes considerable time to obtain a flow equlibrium!
32 DTU,DTU,
Risø Department
Effect of dynamic inflow – numerical actuator
disc simulations – step change in loading
33 DTU,DTU,
Risø Department
Effect of dynamic inflow – numerical actuator
disc simulations – step change in loading
34 DTU,DTU,
Risø Department
Dynamic inflow
 *
NW  R  R
r  k2 r
2
 R  k
 k1 r
1
0
NW FW
k2 -0.4783 -0.4751
k1 0.1025 0.4101
k0 0.6125 1.9210
35 DTU,DTU,
Risø Department
Dynamic inflow model in HAWC2
The Dynamic indflow ins handled by two first order

filters coupled in parallel with weight factors
 
 FW r R   FW
*
 
r
R 
R
V   
 NW r R   NW
*
 
r 1.8R
V, y MAX1  3 induc, y ,0.2
R  V 
V , y MIN1  3 induc, y ,2.0
 V , y   V , y 
The non-dim time constants was found to approximately 2.0 for the far wake
contribution and 0.5 for the near wake part
36 DTU,DTU,
Risø Department
BEM procedure
• For each point in the polar grid
1. Get the local wsp and use local induced wsp from previous timestep/iteration
2. Get the local CT calculated at the two neighboring blades if the local wsp and
induction was used there.
3. Account for Prandtl tiploss
4. Interpolate CT based on azimuthal distance between grid point and the two
blades. Now the local CT is known.
5. Calculate local induction factor a
6. Correct mean level of a based on skew inflow angle.
7. Calculate tangential induction factor am
8. Calculate induced wind speeds axial (1) and tangential
9. Correct for azimuthal variations of axial induction u related to skew inflow
angle incl. axial induction influence..
10. Update u and ut in time based on two first order low pass filters. One for near
wake contribution and one for far wake.
• For each point on the blade

11. Get the induced velocity based on azimuthal interpolation of the two closest
grid points at same radius.
37 DTU,DTU,
Risø Department
Validation results
38 DTU,DTU,
Risø Department
Comparison
39 DTU,DTU,
Risø Department
Special shear (1)
40 DTU,DTU,
Risø Department
Special shear (2)
41 DTU,DTU,
Risø Department
Special shear (3)
42 DTU,DTU,
Risø Department
Wake situation
270 90
43 DTU,DTU,
Risø Department
Examples of limitations of the BEM model
All rotors with the same uniform loading, CT=0.89
For such special rotors a user defined a-CT relation can be used if the
characteristic is known from e.g. CFD actuator disc simulations
44 DTU,DTU,
Risø Department
Aerodynamic input in HAWC2
begin aero ;
nblades 3;
hub_vec hub_center 3 ; rotor rotation vector (normally shaft composant
; directed from pressure to sustion side)
link 1 mbdy_c2_def blade1;
ae_filename ./data/H2_blade_ae.dat ;
pc_filename ./data/H2_blade_pc.dat ;
induction_method 1 ; 0=none, 1=normal
aerocalc_method 1 ; 0=no aerodynamic, 1=with aerodynamic
aerosections 30 ;
ae_sets 1 1 1;
tiploss_method 1 ; 0=none, 1=prandtl
dynstall_method 2 ; 0=none, 1=stig øye method,2=mhh method
end aero ;

2D aerodynamic inputs: the ae file
Aerodynamic properties depend on geometric
characteristics:
– Chord length
– Profile geometry -> steady aerodynamic
forces (complex relation, input: profile
coefficient tables)
Input: aerodynamic layout of the blade.

Given in the ae_filename file:
1 Aerodynamic planform 1. Total number of datasets given in the
1 12
0 3.5 100 1 file
0.5 3.5 100 1 2. Set Id number , Tot. number of points
2.1 3.5 99.9734 1 given in the set
4.9 4.2 74.46957 1
7.7 4.9 48.96571 1 In the following lines:
10.5 5.6 35.71321 1 o Spanwise position [m]
21.7 4.9014 20.58616 1
32.9 3.5489 17.32931 1 o Chord Length [m]
35.7 3.24968 16.76627 1
47.033 1.83467 15.65677 1 o Thickness/Chord ratio [%] (link to
52.26 0.913 14.99 1 prof. coeff. tables)
52.9 0.8 14.7875 1 o Profile Coefficient Set

Profile Coefficients: input file
Profile coefficients collected in the pc_filename file:
1 Airfoil data for the nrel 5 mw turbine

8
1 127 17.0 comments here...
-180.00 0.000 0.0198 0.0000 1. Total number of “sets”
-175.00 0.374 0.0341 0.1880
... ... ... 2. Number of profiles in the first set.
+175.00 0.749 0.0955 0.3770 3. First profile in the first set:
+180.00 0.659 0.2807 0.2747
2 127 21.0 comments here... Index | N rows | t/c profile (link to ae)
-180.00 0.000 0.0198 0.0000
-175.00 0.374 0.0341 0.1880 NB: The t/c in increasing order
... ... ...
+175.00 0.749 0.0955 0.3770
+180.00 0.659 0.2807 0.2747 Then for each profile, steady coefficients
... ... ...
8 127 100 comments here... table:
-180.00 0.000 0.0198 0.0000 o Aoa [deg] (arbitrary spacing)
-175.00 0.374 0.0341 0.1880
... ... ... o Cl
+175.00 0.749 0.0955 0.3770 o Cd
+180.00 0.659 0.2807 0.2747
o Cm

Thank you for
your attention
51 DTU,DTU,
Risø Department
Corrections for yawed inflow – extra info
General case of momentum for rotor, Glauert
  
          
 V  A V  ui V  V  2ui  T  A V  ui  2ui 
T m
Which for axial flow is
 
Tz  A V  ui  2ui , z   Tz  A ud  2ui , z 
Geometry relation

ud  V cos   ui , z  V sin 
2
 
2

2
 ui2, z 2 cos  ui , z 
2  ui , z
 ud  V 1  
2
2  a
V V  V ,local
  
 ui2, z 2 cos  ui , z 

  2ui , z 
2
Tz  A V 1  2

V V 
  
T  12 AV02 4a1  a 2  2a cos   2

1
CT  1
T

 4a 1  a 2  2a cos  y 
1
2
2 AV0
2

HAWC2 Course: Lesson 3: Aerodynamic Modeling and Implementation in Hawc2

Uploaded by

Copyright:

Available Formats

HAWC2 Course: Lesson 3: Aerodynamic Modeling and Implementation in Hawc2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HAWC2 Course: Lesson 3: Aerodynamic Modeling and Implementation in Hawc2

Uploaded by

Copyright:

Available Formats

HAWC2 course

Lesson 3: Aerodynamic modeling and implementation in

Aerodynamic Aerodynamic drag forces

Induction Airfoil section

5 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

– 2D simulations (panel codes, CFD)

– 2D wind tunnel tests

• Limited angle of attack range:

– Extend airfoil data over a larger angle of attack

– Apply 3D corrections: stall delay effects

7 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

Some useful references:

8 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

2D aerodynamic model should account for dynamics (unsteady forces) in:

• Flow separation (dynamic

9 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

10 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

• Clatt and Clfs computed from steady

• Dynamics from weight, separation

11 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

Please use the MHH stall

Increasing Fixed wake

The reaction of the blade forces applied

BEM equations (relation between

Applied before and after rotor disc

ACVV0  A1u1   ( ACV  A1 )V0  m side  side  A1 (V0  u1 )

Similar for the torque

Rotor plane is discretized into

20 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

A non-rotating grid is defined in which induced velocities are

Axial thrust at given

The local thrust

a  k3CT 3  k2CT 2  k1CT  k0

Angular momentum on annular ring

dQ   (2 r dr )(1  a)V (2a' )

Tangential blade forces on annular ring

 tip correction model to account for a finite

 a yaw model to account for skew inflow

 a dynamic induction model to account for

The Prandtl tip correction

Based on model by Glauert

 Is the azimuth angle

In HAWC2 the Coleman method is used

It takes considerable time to obtain a flow equlibrium!

The Dynamic indflow ins handled by two first order

• For each point on the blade

All rotors with the same uniform loading, CT=0.89

45 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

Input: aerodynamic layout of the blade.

46 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

1 Airfoil data for the nrel 5 mw turbine

48 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

T  12 AV02 4a1  a 2  2a cos   2

52 Risø DTU, Technical University of Denmark Presentation name 17/04/2008

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.