model Validation

Wind & Water Controls
61

Dear Jason,

Thanks for your reply.

I have attached the plots of tower top shear force and moment by considering TStart = 60 sec for LC 3.1.

Thanks,
Satish J

LC 3.1 - Comparison tower top moment 60 sec transient.jpg
LC 3.1 - Comparison tower top moment 60 sec transient.jpg840×630 66.9 KB

LC 3.1 - Comparison tower top SS SF 60 sec transient.jpg
LC 3.1 - Comparison tower top SS SF 60 sec transient.jpg840×630 57.3 KB

LC 3.1 - Comparison tower top FA SF 60 sec transient.jpg
LC 3.1 - Comparison tower top FA SF 60 sec transient.jpg840×630 39.9 KB

62

Dear Satish,

It looks like you found the issue–the thrust is different. This offset in thrust times the 110-m moment arm is causing the difference in mean bending moment at the mudline for LC 5.1.

My guess is the difference is in the aerodynamic modeling set up, similar to what was recently discussed in the following forum post: NREL 5MW Reference WT - Power Curve.

Best regards,

63

Dear Jason,

Thanks for your reply.

I went through the post, I have downloaded aerodyn14 from github.com/old-NWTC/AeroDyn14 and then I copied the files in sample to my fast folder and changed the CompAero = 1 and the location of Aerodyn_tower. I am getting the following error,

[code]FAST_InitializeAll:FAST_Init:FAST_ReadPrimaryFile:1/VTK_fps is not an integer multiple of DT.
FAST will output VTK information at 15.385 fps, the closest rate possible.
FAST_InitializeAll:ED_Init:ED_ReadInput:ReadBladeInputs:ReadBladeMeshFileAD: Error allocating
arrays for blade mesh input properties: BldNodes must be at least 1.

FAST encountered an error during module initialization.
Simulation error level: FATAL ERROR

Aborting OpenFAST.[/code]

Is this the correct way of switching to aerodyn14?

Thanks,
Satish J

64

Dear Satish,

AeroDyn v14 is already included in OpenFAST, so, there is no reason to download AeroDyn v14 separately and recompile.

Regardless, the error points to a problem in your input file formatting…I would guess a problem in your AeroDyn v14 input file. A sample AeroDyn v14 input file for the NREL 5-MW baseline turbine is included in the OpenFAST r-test–see: github.com/OpenFAST/r-test/blob … _noTwr.dat.

Best regards,

65

Dear Jason,

Thanks for your reply.

I have copied the file which you suggested for aerodyn14 and setting CompAero = 1. I got the following result,

Thanks,
Satish J

LC 3.1 - Comparison tower top FA SF 60 sec transient with aerodyn14.jpg
LC 3.1 - Comparison tower top FA SF 60 sec transient with aerodyn14.jpg840×630 47.4 KB

66

Dear Satish,

OK, so switching from AeroDyn v15 to AeroDyn v14 definitely brings the results closer. I’m not sure what your goal is–to match exactly) or just understand why you see differences? Regardless, to get the results even closer, would require changes to to the AeroDyn inputs.

Best regards,

67

Dear Jason,

Thanks for your reply.

  1. I would like know to what is the difference between aerodyn14 and aerodyn15 as we can observe significant amount of difference in magnitude. Is it OK to use aerodyn15?

  2. What needs to be changed for aerodyn to get it closer as I can see we need to change blade properties but these properties matches well with the properties in the Definition pdf.

Thanks,
Satish J

68

Dear Satish,

Considering that the original solution from OC3 Phase I is from a much older version of FAST (v6, with AeroDyn v12.6), it is difficult for me to say what input file changes you’d need to make to AeroDyn v14 to more closely match the old results. I wouldn’t change physically known quantities like twist, chord, and airfoil data; instead, you’d have to play around with different wake model options.

There are many differences between AeroDyn v15 and AeroDyn v14, as summarized in Appendix G of the old AeroDyn v15 User’s Guide and Theory Manual: drive.google.com/file/d/14DoKt0 … sp=sharing. I would say generally that AeroDyn v15 is more accurate.

Best regards,

69

Dear Jason,

Thanks for your reply.

I have attached the FA SF at the tower top after playing with the model options (setting StallMod = STEADY, InfModel = DYNIN, IndModel = WAKE) , could I know why there might be the difference in magnitude and what might be the reason for the disturbance in the graph.

Thanks,
Satish J

LC5.1_Task23_Comparison 3 - -20 FA SF Steady Dynin Wake.jpg
LC5.1_Task23_Comparison 3 - -20 FA SF Steady Dynin Wake.jpg840×630 51.8 KB

70

Dear Satish,

I can’t really comment on this result by only looking at this one figure. When trying to understand the FAST output, in general, I would recommend plotting several wind turbine outputs (blade, drivetrain, tower, control motions and loads) and plot PSDs in addition to time series.

Best regards,

71

Dear Jason,

Thanks for your reply.

I have attached the plots in the document, it looks like need to change in Servodyn module, but I am not sure what needs to be changed.

Thanks,
Satish J
Plots.pdf (720 KB)

72

Dear Satish,

I would not say that there is anything wrong in your ServoDyn inputs. It is just that the aerodynamic loads are slightly different, resulting in a bit different response. Again, I’m not really sure what your objective is.

Best regards,

73

Dear Jason,

Thanks for your reply.

I want to match all outputs with referred model. Is there anyway that I can match the aerodynamic loads?

Thanks,
Satish J

74

Dear Satish,

If your objective is to match exactly to the old results, my guess is that you’d need to downgrade to the old version of FAST that was used to produce the OC3 Phase 1 results. But this version is no longer supported by NREL and I would not recommend doing that. Instead, I would recommend accepting that there are differences between the old and new version and that the new solution is generally preferred.

Best regards,

75

Dear Jason,

Thanks for your reply.

Okay, thanks! Other than this, Is there any way that I can validate the model with other?

Thanks,
Satish J

76

Dear Satish,

I would say that you’ve already verified your model. That is, you compared your results to another similar solution and got reasonable agreement, and the differences between the two solutions are explainable (differences in aerodynamic implementation).

Validation of your model would involve checking your solution against physical measurements or against a higher fidelity model (which I’m not sure exist for the NREL 5-MW turbine atop the OC3 monopile).

Best regards,

77

Dear Jason,

Thanks for your reply.

I will have a look, thanks!

Thanks,
Satish J

78

Dear Jason,

Could I get the files of OC4 for verification of my model.

Thanks,
Satish J

79

Dear Satish,

Are you referring to the various simulation results from Phase I and Phase II of the OC4 project? These are available on the following Google drive: drive.google.com/drive/folders/ … EZLdDRxX2s.

Best regards,

80

dear Jason,
im using FAST v8.16 and Aerodyn v15 but i want to switch to Aerodyn v14
i found in some posts here the input file for Aerodyn v14 and i changed compaero to 1 in FAST v8.16 input file but im receiving the following

FAST_InitializeAll:ED_Init:ED_ReadInput:ReadBladeInputs:ReadBladeMeshFileAD:Invalid numerical
input for file “.\5MW_Baseline/NRELOffshrBsline5MW_OC3Hywind_AeroDyn.dat” occurred while trying
to read BldNodes.
Do i have to make more changes or i need more files for Aerodyn v14 that dont exist in FASTv8.16 archive that i have downloaded?