Static structural test on rotor blade comparing BeamDyn and ElastoDyn

Hello everyone,
I am a student and I have been working on a term paper for the past months. My goal is to simulate a static test on a rotor blade that is put in the 3 o’clock position and pulled towards the ground. I am using Structural Control to apply the prescribed force on the blade. Then I want to compare the deflection, rotation, sectional forces and moments calculated by BeamDyn and ElastoDyn.

• I am unsing the 5MW-Baseline turbine

1 CompElast/
2 CompElast
1 CompServo

ElastoDyn:
False All DEGREES OF FREEDOM except for the tower and blade bending modes
49 BldNodes
0 BlPitch(1)
90 Azimuth

BeamDyn
49 station_total

ServoDyn/Structural Control
1 NumBStC
4 StC_DOF_MODE
45 StC_P_Z
1 PrescribedForcesCoord
Fz -1.0e5 (pulling the horizontal blade towards the ground)

Now, I have a few questions regarding my results.

1. I have read in an old post, that there was a bug in BeamDyn calculating the local forces and moments. Has that bug been fixed and can I assume all deflections, moments and forces that are put out by BeamDyn and ElastoDyn to be correct?

2. I plotted BeamDyn’s B1N0xx_Fyr and ElastoDyn’s B1N0xxFLy over the beam length (assuming no gravity). From Mechanics I would expect the local force FY to be constant over the beam length, up until the point where the prescribed force hits the beam. There should be no force in the free end of the cantilevered beam, much like this:

BeamDyn calculates a different outcome with the local force still quite high after the point where the prescribed force is applied (dashed line)

Could this have to do with the torsion of the blade or could it be a bug in BeamDyn’s calculations?

3. For the local force Fx however, I can not explain ElastoDyn’s outcome (plotting B1N0xx_Fxr and B1N0xxFLx). The pitch angle and the direction of the prescribed force are still the same. Do you think I have done a mistake with my simulation?

Thank you very much and best regards
Laura

Dear @Laura.Goetza,

I’m not aware of any existing bugs in the outputs of BeamDyn in OpenFAST v3.5.0.

You say “assuming no gravity”; did you actually set `Gravity` = 0?

I agree that torsion may be playing a role in BeamDyn. I also think that the structural pretwist of the NREL 5-MW baseline blade is playing a role in both the BeamDyn and ElastoDyn results, especially considering that the ElastoDyn outputs you are using are expressed in a local blade coordinate system that is aligned with the structural twist. To gain more confidence in the results, I would suggest running these tests with a much simpler beam with uniform properties along its length and no structural pretwist.

Best regards,

Dear @Jason.Jonkman,

Thank you for your reply. I followed your suggestion and used a simpler beam with no twist, homogeneous properties and the same bending stiffnesses in edge and flap direction.

I found out that the sectional force along the local xb-axis in ElastoDyn (B1N0xxFLy) has to do with the structural twist. If I set it to 0° along the blade, I get the outcome that I would expect (like I mentioned in my example above).

With BeamDyn however, even with the simple beam, I get the same outcome as before and BeamDyns FxL and Fxr look the same. And yes I set Gravity = 0.

In my understanding (and please correct me if I’m wrong) there should not be any sectional force in the beam past the point where the external force is applied, like we see in ElastoDyn’s outcome. In BeamDyn, the force decreases more gradually. Do you have an idea how this outcome could be explained with the way BeamDyn calculates the sectional forces?

I also have a question about the local axes in ElastoDyn. Is there a local axis at each node (BldNodes) in ED, like there is a local blade coordinate system at each key point in BeamDyn?

Dear @Laura.Goetza,

I’m glad to hear that the ElastoDyn solution makes sense now.

Can you clarify how you are getting a static solution? Is this the solution after all start-up transients die out?

For BeamDyn, can you clarify what element order (`order_elem`) is being used? Do the results change if you increase the order?

Best regards,

To get a static solution, I simulate over 80 seconds until the results seem stable and then I take the mean value over the last 10 seconds for each output parameter. My order_elem was 5 and I set it to 8 but the sectional force still decreases in the same way.

With this drawing, I think I can explain to myself how the sectional force expressed in l (FyL) decreases gradually, as the blade deflects. Because the local blade coordinate system is oriented and tilted with the blade axis. So the sectional force is divided into a FyL- and a FzL-component and FyL therefore decreases gradually.

But I can not explain why the sectional force expressed in r looks like this.