Modelling Blade Damage

Discuss theory and modeling of wind-turbine structures.

Moderator: Bonnie.Jonkman

Mark.Capellaro
Posts: 12
Joined: Wed Aug 10, 2016 7:50 am
Organization: Disgruntled Goat GmbH
Location: Bavaria

Modelling Blade Damage

Postby Mark.Capellaro » Fri Aug 12, 2016 1:38 am

Hello Fast/Structural Analysis Forum,

I am trying to model damage in a single blade by reducing the flap bending stiffness in a single blade (as a first test).

I am running Fast v8.16 and using the NREL5MW onshore as the baseline turbine model. Basically cut and paste from the test26.

Proposed but as of yet unsucessful method:
1. in the _ElastoDyn_BDoutputs.dat file make blade 3 a separate BldFile "_Blade_3.dat" BldFile(3)
2. In "_Blade_3.dat" significantly reduce flap stiffness at one location (as a test)
3. in the "_BeamDyn.dat" BDBldFile(3) create a new _BeamDyn_3.dat file and reduce the same parameter (flap EI) the same amount (and ask myself which of the inputs is actually used)

The simulation results (8ms turbulent wind) showed no noticeable differences between blade 1 and 3 loads and deflections.

A later attempt with the input: AdjFlSt - Factor to adjust blade flap stiffness (-) in the "_Blade_3.dat" file also did not change the results.

So is this something I am doing wrong or are blade structures/properties in FAST unitary?

Mark C

Jason.Jonkman
Posts: 5071
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Modelling Blade Damage

Postby Jason.Jonkman » Fri Aug 12, 2016 10:59 am

Dear Mark,

FAST can consider structural differences between the various blades.

I wouldn't expect any difference in the results for your cases (1) and (2) and the change to AdjFlSt, but I would expect differences in your case (3).

In cases (1) and (2), it appears that you are changing the stiffness of the blade in ElastoDyn. However, with BeamDyn enabled (by setting CompElast=2) in Test 26, the blade-related inputs and outputs from the ElastoDyn module are unused, replaced with those available in the BeamDyn module (as explained in the FAST v8 ReadMe file). That is, if CompElast=2, ElastoDyn inputs FlapDOF1, FlapDOF2, EdgeDOF, OoPDefl, IPDefl, TipRad, TipMass(1-3), BldNodes, BldFile(1-3), NblGages, and BldGagNd are unused and ElastoDyn outputs for the blade tip motions, blade local span motions, blade root loads, blade local span loads, and internal blade mode DOFs are unused.

For case (3), it appears that you are changing the blade stiffness of only the 3rd BeamDyn blade, which I would expect to have an influence on the results, as long as you are reviewing pertinent outputs.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Mark.Capellaro
Posts: 12
Joined: Wed Aug 10, 2016 7:50 am
Organization: Disgruntled Goat GmbH
Location: Bavaria

Re: Modelling Blade Damage

Postby Mark.Capellaro » Wed Aug 17, 2016 9:21 am

Thanks Jason,
Thought I had set the BeamDyn on. I increased the area of decreased flap stiffness and now the results show a difference in blade deflection.
Mark C

Mark.Capellaro
Posts: 12
Joined: Wed Aug 10, 2016 7:50 am
Organization: Disgruntled Goat GmbH
Location: Bavaria

Re: Modelling Blade Damage

Postby Mark.Capellaro » Mon Nov 14, 2016 2:45 am

Hello Forum, I am modeling the NREL 5MW with FAST8 using the Test26 as the basis.

I added some accelerometers to the NREL 5MW model in Test26 (BeamDyn input, 3 NNodeOuts with "N1TAXg, N2TAXg, N3TAXg"
"N1TAYg, N2TAYg, N3TAYg" added to the OutList).

The results at the root show a strange 1P Y direction spike:
1P Spike.png
1P Spike.png (35.71 KiB) Viewed 1969 times

that I can not explain.
The nodes shown above (if I understand correctly) where the signal (acceleration) is computed should be at the blade root.

All blades have at the same time the distorsion (so no tower shadow).
The distorsion is much less but still visible at the second node (accelerometer) and can be seen on the root accelerometer X direction signal also.

I first tried to simplify the wind input but a steady wind has the same result.
(other attempts to reverse engineer have included changing the time step and the order elem for the blade, neither of which improved the situation).

Anyone want to try to explain what is happening at the root to make these signals?

Jason.Jonkman
Posts: 5071
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Modelling Blade Damage

Postby Jason.Jonkman » Mon Nov 14, 2016 10:25 am

Dear Mark,

I am able to reproduce your problem running Test26 using FAST v8.16 compiled in single precision. However, using the double precision version solves the problem. We typically recommend using the double precision version of FAST when running models with BeamDyn due to numerical sensitivities in BeamDyn that we are still trying to figure out.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Mark.Capellaro
Posts: 12
Joined: Wed Aug 10, 2016 7:50 am
Organization: Disgruntled Goat GmbH
Location: Bavaria

Re: Modelling Blade Damage

Postby Mark.Capellaro » Tue Nov 15, 2016 3:02 am

Thanks Jason


Return to “Structural Analysis”

Who is online

Users browsing this forum: No registered users and 1 guest