OpenFast vs Bladed Model

Wind & Water Computer-Aided Engineering Software Tools
1

Dear @Jason.Jonkman,

I am trying to translate a 2.0MW model from Bladed to OpenFast. For this, I am currently running simulations with steady wind for both, Bladed and OpenFast models to ensure that are responding similarly.

I am comparing three curves:
1- Bladed model with generic controller, tipically used in Bladed.
2- Bladed model with OpenDiscon controller.
3- OpenFast model with OpenDiscon controller.

1&2 are giving the same values, so we discarded that the problem is in the controller differences. However, when using the OpenFast model, there is an approximate shift of 0.5 m/s, although the overall trend until rated wind speed seems to match:

PowerCurve
PowerCurve2539×1694 132 KB

And the rotor speed is also below the Bladed model

Here are some ideas considered to explain these differences:

-Could it be that I am not properly accounting for the free wind speed in OpenFast? The variable I am using for this is Wind1VelX.

-Is there an issue with the Kopt? I have made modifications here, but there hasn’t been any significant improvement.

-Am I incorrectly defining the so-called “Blade Set Angle” in Bladed when transferring to OpenFast? I understand there isn’t a direct variable for this in OpenFast, so I modified the Aerodynamic Twist distribution to incorporate the Blade Set Angle. However, this modification hasn’t released better results.

Am I overlooking something? Any insight you could provide would be greatly appreciated.

Thank you in advance.

2

Dear @Ainhoa.Alberdi,

Just a few comments:

  • InflowWind module output Wind1VelX is the ambient wind speed at the first point specified in list of coordinates in InflowWind.
  • I’m not familiar with Bladed’s definition of blade set angle; are you referring to the blade-pitch angle? In OpenFAST, the orientation between the chordline and rotor plane is pitch + twist.
  • Before enabling the controller, I would first verify that the aerodynamic loads (thrust and torque) are the same between OpenFAST and Bladed for the same wind speed and rotor speed. If they are not, Kopt would have to differ between the two, but likely that means there is some other error in the model conversion. There could also be small differences in the aerodynamics models between Bladed and OpenFAST, but I would not expect large differences.

Best regards,

3

Dear @Jason.Jonkman ,

As commented yesterday, the rotor speed we are finding with OpenFast is lower at wind speeds before rated:

RotSpeed
RotSpeed1919×1198 78.3 KB

So specially the aerodynamic torque we are obtaining with OpenFast is 16% lower approx.

Any insight you could provide would be greatly appreciated.

Thank you in advance,

4

Dear @Ainhoa.Alberdi,

I would start by simplifying the model as much as possible–How much difference in aerodynamic torque between Bladed and OpenFAST do you see for the same rotor speed if you disable all structural degrees of freedom in both models?

Best regards

5

Dear @Jason.Jonkman ,

Thank you very much for your indications.

I am currently following this approach, but something really odd is happening with the polar curves. I have obtained Cl, Cd and angle of attack for 10m/s at different blade sections. For example, at section 24m from the blade root, I am obtaining the next values:

AoA: 3.78deg
Cl: 0.711
Cd: 0.008

This section is linked to airfoil number 4. If I check this airfoils Cl and Cd coefficients for that Angle of Attack Range, I have this values:

Alpha Cl Cd
3.5 0.951 0.009
4 1.012 0.009

How is it possible to have such a low lift coefficient if is not happening between the angles of attack that is interpolating?

Thank you again, really appeciated.

6

Dear @Ainhoa.Alberdi,

It can happen that the lift and drag don’t match the static airfoil data at a given angle of attack if you have unsteady airfoil aerodynamics enabled (UA_Mod > 0 in AeroDyn, i.e, the airfoil data follows a hysteresis loop around the static polar data). It could also be that you have multiple airfoil tables for a given airfoil and you are interpolating based on Reynolds number or control (AFTabMod > 1 in AeroDyn).

Best regards,