model Validation

Hi,
I want to validate my modeling with FAST but i dont know how can i do that.
Please, can any one help me to do that?

Thank a lot,

Hello Vahid,

you should probably be more pecise. What would you like to validate? Your model setup, your simulation results, …?

Hi Paul,
Im sorry i forgot tell about my device.
I modeled 660 Kw Vestas wind turbine. So, i want validate my modeling with FAST model. Actually, i want validate my simulation results.

Thanks a lot,

Dear Vahid,

I agree with Paul. Please clarify your question. From where you getting the data needed to validate your model? Under what conditions? What quantities of interest are you interested in validating? What’s your validation metric?

Best regards,

Dear Jason,
Thank you very much for your quick reply,
I will send you more details.
Kind regards,

Hi,

I have simulated the 5MW_OC3Mnpl_DLL_WTurb_WavesIrr.fst model in openfast corresponding to the data provided in nrel.gov/docs/fy15osti/63067.pdf. But I got the following results but doesn’t match well with the result obtained in the paper corresponding to monopile and the load case here I am referring is 5.1 and I have attached the plots of Time vs OoPDefl1 and Time vs RootMyc1.

And is there any reference where I can validate the openfast model in terms of whether the model is aligned or not.

Thanks,
Satish J
Comparison FAST.docx (86.5 KB)

Dear Satish,

Actually, it looks like your results are quite close, although there is a small difference in the mean values. I’m not aware that the NREL model and results from the paper you referenced were made available, although the model of the NREL 5-MW turbine atop the monopile from the FAST v8 CertTest and OpenFAST r-test are likely quite close to the model that was used in this paper.

Best regards,

Dear Jason,

Thanks for your reply.

Does that mean the result obtained using openfast are correct?

Or how to validate the model (openfast) in terms of whether the output is correct?

Thanks,
Satish J

Dear Satish,

Yes, I believe your answers are correct. I’m not sure what is causing the small difference in mean values (perhaps a different mean wind speed?), but overall the trends look correct.

If you want to plot your results against other OC3 participants, all results generated during the OC3 project are available on Google drive–see: drive.google.com/drive/folders/ … 3RkZ3FHVlE (the FAST v8 results from the paper you referenced are not included here because these were generated after the OC3 project concluded, but the plots in that paper were made by comparing to the results stored in this Google drive).

Best regards,

Dear Jason,

Thanks for your reply.
I will check.

Thanks,
Satish J

Dear Jason,

I run the 5MW_OC3Mnpl_DLL_WTurb_WavesIrr.fst model using Openfast and thought of comparing the result which was obtained in the document which you sent for the load case 5.1 - Phase I Monopile with Rigid Foundation. The following are the results and looks like my simulation results are wrong. Please could you advice me on this and I have attached the relevant files and plots.

Thanks,
Satish J
NRELOffshrBsline5MW_OC3Monopile_HydroDyn.rtf (18.1 KB)
NRELOffshrBsline5MW_OC3Monopile_ElastoDyn.rtf (13 KB)
5MW_OC3Mnpl_DLL_WTurb_WavesIrr.out.rtf (953 KB)

Dear Jason,

Continued…I couldn’t able to add more than 3 files in my previous post. So I am attaching the result and fast file.

Thanks,
Satish J



5MW_OC3Mnpl_DLL_WTurb_WavesIrr.fst.rtf (7.24 KB)

Dear Satish,

Can you clarify what you are comparing your results too such that you believe your results to be wrong? Which figure from which document are you comparing to?

Best regards,

Dear Jason,

Thanks for your reply.

I am comparing with the results obtained in the document which you sent earlier - drive.google.com/drive/folders/ … 3RkZ3FHVlE (Document location - Task 23 OC3\Phase I Monopile with Rigid Foundation\Simulation Results\Rev 5 - Plus New Comers - Updated\SimulationResults_NREL_FAST). Load case considered for comparison is 5.1 (Task 23 OC3\Phase I Monopile with Rigid Foundation\Simulation Results\Rev 5 - Plus New Comers - Updated\LoadCaseMatrix_PhaseI_070112).

I plotted Time vs wave elevation and Time vs fore-aft moment at the mudline from results obtained in the above document and then I compared with simulation results which I have attached in the previous post.

Thanks,
Satish J

Dear Satish,

OK, thanks for clarifying.

One big difference I see is that the original OC3 Phase I results for load case 5.1 assumed wave linearity (Airy waves). However, I see that you have enabled second-order wave-kinematics (WvDiffQTF = WvSumQTF = TRUE) in your analysis (hence why your wave elevation is not sinusoidal). The second differences is in regards to wave stretching. The original OC3 Phase I results included Wheeler stretching for the wave kinematics, however, this functionality is not available in the version of OpenFAST that you are running. The lack of Wheeler stretching will impact the wave loads and resulting bending moment at the mudline.

Best regards,

Dear Jason,

Thanks for your reply.

Yes. But fore-aft bending moment is decreasing with as I have consider the wave properties to be constant throughout the simulation time (hence it should not decrease right?) and what might be the solution for this and wave elevation?

Thanks,
Satish J

Dear Satish,

I suppose you are referring to the start-up transient in the mudline moment (and other outputs) that occurs over the first 30 s or so of the simulation? These happen in all simulations and can be reduced by choosing initial conditions that are more line with their expected values for a given simulation. Regardless, the start-up transients were eliminated in the OC3 Phase I results, i.e., by not outputting data until after the start-up transient has ended (e.g., by setting TStart > 30 s).

Best regards,

Dear Jason,

Thanks for your reply.

I have attached servodyn file where I have considered TStart as 30sec (but it says unused).

  • What might be the case in the wave elevation?

  • Do I get the input files used for phase 1 load case 5.1?

Thanks,
Satish J
NRELOffshrBsline5MW_OC3Monopile_ServoDyn.rtf (11.4 KB)

Dear Satish,

TStart is an input parameter in the OpenFAST primary (*.fst) file, not in ServoDyn.

I don’t really understand your other questions.

Best regards,

Dear Jason,

Thanks for your reply.

Following the above quoted post, how can I get the wave elevation sinusoidal?

Thanks,
Satish J