WindFarm_Wake

[oguzcrs11]

Dear all;

My name is Oğuzhan, and I am currently a Master's student in Turkey. I have been working on analyzing fatigue for downstream turbines, particularly focusing on accounting for wake effects. However, due to my limited experience with OpenFAST, I have been progressing step by step and have encountered some challenges.

From my research, I understand that simulation data and outputs for the NREL 5MW turbine have been used for fatigue analysis with MLife. However, I have not been able to find exercises or datasets that involve multiple turbines. At this point, I am looking for a wind farm dataset that includes wake effects for downstream turbines, which I can use in QBlade or Ashes for further analysis or "it might be more accurate to say wind data".

As I am still learning how to connect the information I have gathered, any guidance or resources you could provide would be invaluable. Additionally, I plan to use QBlade or Ashes for visualization since I am not fully familiar with Paraview.

Please let me know if any part of my request is unclear or if there are specific details you need. I greatly appreciate your time and assistance.

Thank you for your consideration.

Best regards,

[jjonkman]

Dear @oguzcrs11,

I'm not sure I fully understand what you are asking, but it sounds like you want turbulent inflow for a given wind turbine that takes into consideration not only ambient inflow turbulence, but also the turbulence from upstream wind turbine wakes.

The IEC design standard 61400-1 outlines a simple "effective turbulence" model that could be applied to this, which works well on average, but misses details associated with the wind farm layout and operational conditions.

For the later, tools such as FAST.Farm can be used to model wind turbine aeroelastics in OpenFAST combined with farm-level wake and array effects from multiple turbines in the farm. But if you were applying FAST.Farm, then you would be using OpenFAST rather than QBlade or Ashes.

FYI: We recently wrote a paper comparing FAST.Farm against the IEC effective turbulence model, which may be informative to you: https://wes.copernicus.org/articles/8/1475/2023/.

Best regards,

[jjonkman]

Dear @oguzcrs11,

In ElastoDyn, the blade structural analysis nodes are placed automatically without additional user input. That is, the blade (between the root and tip) is discretized into BldNodes equal-length elements and the blade structural analysis nodes are located in the centers of those elements such that the first node is located at 0.5*(TipRad-HubRad)/BldNodes + HubRad along the pitch axis from the hub centerline, the second node is located at 1.5*(TipRad-HubRad)/BldNodes + HubRad, etc. for each node from 1 to BldNodes. These blade structural analysis node locations are written to the ElastoDyn summary (*.ED.sum) file.

In AeroDyn, the blade aerodynamic analysis nodes are user-located and these nodes can be placed independently from the structural analysis nodes.

Best regards,

[oguzcrs11]

Dear @jjonkman,

If I'm not mistaken, BEMT is extension of Actuator disk and many actuator method is in the literature, Is it possible to add another method instead of BEMT in OpenFast e.g. actuator line.
Please excuse me if my question is unnecessary. Just I want to compare the results of different methods.

Best Regards.

[jjonkman]

Dear @oguzcrs11,

I agree that the "momentum" part of BEMT relies on actuator disk assumptions, although the "element" part of BEMT relies on an actuator line.

OpenFAST itself can already bypass actuator disk assumptions by switching from BEMT to the free-vortex wake method (in the OLAF submodel of AeroDyn).

Moreover, OpenFAST has also been coupled to high-fidelity CFD through an actuator-line formulation (e.g., in OpenFOAM (SOWFA) and AMR-Wind) and through a geometry-resolved formulation (e.g,. in NaluWind).

Best regards,

[oguzcrs11]

Dear @jjonkman,

Is it possible to use cfd in Fast or should I use SOWFA for it. I would like to add a chapter in my thesis a title like that "comparsion of the blade element momentum with cfd results" for single turbine.

Best Regards.

[jjonkman]

Dear @oguzcrs11,

High-fidelity CFD is not directly available within OpenFAST, but is available through an actuator-line formulation (e.g., in the coupling between OpenFAST and OpenFOAM (SOWFA) or AMR-Wind) and through a geometry-resolved formulation (e.g,. in the coupling between OpenFAST and NaluWind).

Best regards,

[oguzcrs11]

Dear @jjonkman,

All your reply was very helpful for me thanks a lot.
I don't know how necessary it is to compare elastodyn and beamdyn results, but I saw that there was an article and frequent questions about this in the past. based on your past answers I applied condition of numcrctn:1 and I started to study similar results between two results with this condition. In NREL form your reply was "I would compare the deflections using the “xb” axis in ElastoDyn against the “xr” axis in BeamDyn." and I printed this values on chart but I didn't understand why for beamdyn result showed 0

e.g. RootFxb1-B1RootFxr

all beamdyn result =0 for root forces and moments

Best Regards.

[jjonkman]

Dear @oguzcrs11,

Can you clarify what you are selecting for the blade root outputs from BeamDyn in BeamDyn's OutList? Does the OpenFAST output file say that the blade root outputs from BeamDyn you've selected are "INVALID" (in which case OpenFAST will output zeros)?

Best regards,

[oguzcrs11]

I selected the variables under the heading of reaction loads according to the outlistparameters in the excel file;

everything else is the same with rtest

Is there a chance it's related to fast dt?

Best Regards.

[jjonkman]

In the OpenFAST output (*.out) file, are the units of these blade-root loads showing up correctly, or are they listed as "INVALID"? Are you sure you are grabbing the correct column of data in the output file?

Best regards,

[oguzcrs11]

I never thought to look at the invalid units in the out file but why, Did I write the wrong variables but these are same with at excell
Is there a step you can suggest?

Best Regards.

[jjonkman]

My guess is that you are grabbing the ElastoDyn outputs (RootFxb1, etc.), which are invalid, rather than the BeamDyn outputs (B1RootFxr, etc.).

Best regards,