15MW airfoil sections and aerodynamic reference points

Wind & Water
1

Dear all.

In reference to the Aerodyn model and turbine definition data in the repository

and the definition document I would like to ask the following:

How are the geometry and the polars of the 50 aerofoil sections in the Aerodyn model derived from the 8 ‘mother’ aerofoil sections described in the definition document of the wind turbine?

And furthermore, why is the reference point for the polars not the same for all the sections?
I understand this is an arbitrary point but it still interests me to know why it is not chosen as e.g. 0.25 for all the aerofoil sections? Is the reference point the theoretical Aerodynamic Centre: the location with respect to which the pitching moment does not change for small changes in the angle of attack?

Thank you in advance,
Duncan

2

Hello Duncan,
thanks for your email. The interpolation of airfoil polars and coordinates is performed by this WISDEM function

As briefly described here Blade — windIO windIO documentation, in between the user-defined airfoils, the interpolation scheme for both coordinates and polars is the Piecewise Cubic Hermite Interpolating Polynomial (PCHIP), which is implemented in Matlab(Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) - MATLAB pchip) and in the Python library SciPy(scipy.interpolate.PchipInterpolator — SciPy v1.9.3 Manual).
The chordwise position of the aerodynamic center is also interpolated with a chip and is 50% at blade root (cylinder) and 25% for the other airfoils.
I hope this helps.
Best regards,
Pietro Bortolotti

3

Dear Pietro,
Thank you for answering my questions.
If I understand correctly the thickness/chord is interpolated from the thickness/chord of the ‘parent’ sections at their appropriate span wise stations to a full span wise distribution of the thickness/chord.

Next the aerofoil properties are interpolated based on the local thickness/chord value at a particular span wise station.

Thank you again and best,
Duncan

4

Dear @Pietro.Bortolotti

First of all, I appreciate your detailed description above.

Currently, I am trying to calculate the IEA 15MW power curve using an eroded blade airfoil.

Unlike the NREL 5MW blade, the IEA 15MW blade uses 50 different airfoil polars interpolated by the mother airfoil.

So I calculated several new mother airfoil data sets (Cl, Cd) based on the FFA series (used in the IEA 15MW wind turbine blade) using CFD and want to replace the original 50 airfoil polar data sets with my own.

Here are my questions.

  • If I understand your answer correctly, should I interpolate my airfoil data (Cl and Cd values) using PCHIP in Python or Matlab?
  • Why are mother airfoil positions not included in the 50 interpolated airfoil positions in the IEA-15-240-RWT.yaml file?

I haven’t calculated PCHIP, so I would sincerely appreciate it if you gave me any small tips.

Best regards,
Sangwon

5

Hello Sangwon,
the link that I shared previously in this conversation is still valid. That’s the section of the code that we use at NREL to interpolate the master airfoils specified in the yaml to the n (50) stations used in OpenFAST. We use PCHIP from the SciPy library. We don’t include 50 airfoil datasets in the yaml because usually that’s not what a rotor designer has available. Rotor designers design and test a few master airfoils, which then get interpolated. Note that you should use the PCHIP interpolator if you want to match our exact interpolation scheme, but there is a number of interpolators out there that you could choose to use and find more appropriate for your scope.
I hope this helps

1 Like
6

Dear @Pietro.Bortolotti

I appreciate your detailed answer, and it helped a lot!

I found 53 stations in the yaml file, but OpenFAST uses 50 stations.

It looks like the blade tip section has been subdivided.

Is there any reason for this?

Best regards,
Sangwon

7

The team has been working with different meshes over the years. 50 is a good start, but you will certainly see differences if you refine the element size (but computational costs will go up!)

1 Like
8

Dear @Pietro.Bortolotti

I get it.

Thanks again for your kind answer, even to my simple question.

Have a good one!

Best regards,
Sangwon