Local angle of attack computation

Wind & Water Rotor Aerodynamics
1

Dear all,

I am trying to compute the local angle of attack along the blade with OpenFAST under the following conditions:

  • Steady inflow at 12 m/s
  • Wind shear exponent 0.2
  • Blade and tower deformability (using ElastoDyn + BeamDyn)
  • Rotor uptilt (5 deg)

I tried the simulation with steady aerodynamics and usteady one (Minnema variant) but I get something I cannot interpret actually.
With the steady model everything goes as expected: after a short transient I get a periodic solution with a frequency equal to the RPM of rotor, as show in the first picture.

If I use the unsteady model instead, I get a very strange periodic solution which is a combination of at least 3 periodic signals with different frequencies (only one of them seems to match the rotor RPM), see the second picture. Moreover, also the oscillation amplitude is quite different from the steady one.
Using the Gonzalez variant of the unsteady model leads to more proper results even if the oscillation amplitude and mean incidence are different from the steady results, see third picture.
Do you have any guess on why I get this strange result? What do you think is the most reliable result?

Best regards,
Matteo



2

Dear Matteo,

Am I correct to say that:

  • Figure 1 shows the results with AFAeroMod = 1,
  • Figure 2 shows the results with AFAeroMod = 2 and UAMod = 3, and
  • Figure 3 shows the results with AFAeroMod = 2 and UAMod = 2?

Can you clarify which OpenFAST model you are running? Is this a model NREL has provided or one that you have developed yourself?

The accuracy of the unsteady airfoil aerodynamic (UA) models will depend on the settings of the UA model parameters in the airfoil data input file. Again, were these provided by NREL or something you set yourself?

It may help to plot the UA hysteresis loops (lift, drag, pitching moment as a function of angle of attack) for the case with a rigid wind turbine (all structural DOFs disabled) to check if they are as you expect them to be.

Best regards,

3

Dear Jonkman,

Yes, exactly.

I am using the 5MW reference model by NREL.

I used the default values provided by NREL.

I also computed the hysteresis plot and both seem good. The firse one is with UAMode = 2 and the second one for UAMode = 3.
Do you have any clue on what’s the correct one?

Regards,
Matteo

1280×624 60 KB

1280×624 57 KB

4

Dear Matteo,

OK, thanks for clarifying.

I have more experience with UAMod = 3 and little experience with UAMod = 2, but in your case, it looks like UAMod = 2 and UAMod = 3 are very similar (some differences are seen in Cd and Cm, but the values are quite small).

Best regards,

5

Dear Jonkman,

I also think the results are very similar and both seems correct. The main difference remains in the time histories when I consider the structural DOFs (as seen in the first plots I posted).
At this point I think I will proceed with UAMode 2 since it gives a meaningful time history (with values similar to steady model), is it a good idea?

Best,
Matteo

6

Sounds reasonable.

Best regards,