Dear Bonnie,
I am trying to set up a FAST model of the Levenmouth wind turbine, and compare the results with the commissioning results.
I have produced a *.wnd and *.bts wind file with TurbSim64 to use it in Bladed and FAST respectively.
After running the simulation I have realised that the signal produced is not the same in Bladed and FAST (See p1.png).
Both wind files were created with the same input file, and I have checked the turbulence parameters of the wind in both Bladed and FAST, and they are the same.
Should not be the wind signals identic? If yes, have you any idea of what I am missing? If not, which is the cause of having different signals?
Once I have got a result for a certain DLC, I run the simulation with different seeds. I have realised that all the results are very similar i.e. there are not significant differences in the maximum loads. The commissioning results show big discrepancies between the results from different seeds i.e. Loads 50% bigger in the same simulation with different seeds.
Does it make any sense for you?
I have used consecutive seed numbers, could that explain this behaviour?
Thank you in advance!
Kind Regards,
Jordi
Dear Jordi,
I would expect the *.wnd and *.bts file to contain the same information, so, the problem is likely in differences in how the Bladed and FAST models have been set up. Here are a few questions:
*Have you eliminated the shear from Bladed? (which is already included in the *.wnd-generated TurbSim file)
*Are the results simply time shifted between FAST and Bladed? (I’m not sure where Bladed chooses to start using the wind data, but the origin is likely different than FAST.)
*Are you looking at the same point (X,Y,Z) in space?
Certainly different seeds will produce peak loads. And using consecutive seeds is generally OK. I’m not sure I can comment on why the commissioning results are different from your results.
Best regards,
Dear Jason,
Thank you for your comments.
Yes, I have eliminated the shear in Bladed.
No, the results are not merely time shifted. There are also small differences in maximum and minimum amplitudes.
I am not sure about the position. Certainly, I can check the Z (RefHt), but how can I check X and Y in FAST? Is it something to do with the Latitude? Is it the inertial X and Y direction (InflowWind)? If so, yes, they are the same.
Have you any other thoughts about it?
Regarding the peak loads, I guess my problem is related to the number of experiments. I have tested it only for six different seeds. I am going to try with twelve different seeds.
All the Best,
Jordi
Dear Jordi,
Regarding the point in space (X,Y,Z) of the wind output, I’m not sure how Bladed works, but in FAST v8, the (X,Y,Z) location(s) of the wind output are specified via InflowWind input parameters WindVxiList, WindVyiList, and WindVziList.
Another thing to consider is the role of TurbSim input parameter Clockwise on the Bladed results. This parameter will influence the results from Bladed, but not FAST. See the TurbSim User’s Guide for more information.
Best Regards,
Dear Jason,
For the comparsion between FAST8, FAST7 and GH Bladed, I have done it with the same DLL controller, ( FAST 7 controller as subroutine ) at 10 m/s turbulent wind. This is the result, where the down part in the graph shows the pitch angle and the upper part shows the generator speed (Direct Drive) and the middle part shows the wind speed.
For Pitch angle and Generator speed ( Blue is FAST8 - Black is FAST 7- Red is GH Bladed )
For wind speed ( Blue is FAST 8 and Green Is GH Bladed )
My Question is: Please, Give me a comment on this result and Why the pitch angle in FAST is higher than the one in Bladed?
Best Regards
Taha
Dear Taha,
The solutions between FAST and GH Bladed could be different for a number of reasons, including different model setup/features enabled and different theory basis (especially aerodynamic).
I’m a bit more concerned that the FAST v7 and FAST v8 solutions are so different. Normally the FAST v7 and FAST v8 solutions should be very similar to each other if the models are set up consistently. Are you expecting these differences based on the model setup/features enabled between the FAST v7 and FAST v8 models?
Best regards,
Dear Jason
yes, there were few wrong vlaues for setting up the model at FAST7 but I am much concerend about the comparsion between FAST8 and Bladed as you see here. I don’t know why the pitch angel with FAST8 is higher than the one in bladed by almost 1 degree ( In case of pitching ) as shown in the plot down where the red is for Bladed and the blue is for FAST.
For the load comparison, why there is a shift in the edgewise moment as shown down?
- I am also not able to understand why the flap moment is a bit higher for FAST8 in few time slot although the pitch angle for FAST is a bit higer or equal?
I have also few questions for setting up the two models :-
1- For the Generator mass, it is not excluded from the nacelle mass, Right? because in Bladed it is excluded.
2- For the DOF in FAST8, there are 6 DOF for the flapwise bending mode ( 2 for each blade ) and only 3 DOF for the edgewise bending mode ( 1 for each blade), where in bladed there are 6 In plane rotor modes and 6 out of plane rotor modes and for each mode there is a damping factor, is this dapmping factor that is defined in the blade input file as (Blade flap mode #1 structural damping in percent of critical)?, if yes.what is the conversion factor between FAST and Bladed?
3- For the blade Geometry (BlCrvAC, BlSwpAC, BlCrvAng) what could be the corresponding parameters for them in bladed, I have in Bladed those Parameters (Thickness %, Pitch Axis %, Pre-Bend m) which I am not able to see what are the corresponding parameters for them in FAST?
4- StrcTwst is the Twist of the blade which is defined in bladed as Twist?
5-Is TwrCd Aerodynamic drag coefficient for the tower?
Best Regards,
Taha
Dear Taha,
The slightly higher pitch angles in FAST relative to Bladed seem to be correlated with the slightly higher rotor speeds in FAST. Another possibility is that the blade torsion is modeled rigidly in the ElastoDyn module of FAST, but perhaps Bladed is showing some small amount of blade twisting resulting in less pitch required through control?
Could the mean offset in the edgewise moments between FAST and Bladed be related to a difference in coordinate systems?
I’m not familiar with the detailed input/output conventions in Bladed, so, can’t really answer associated questions. I suggest you reach out to Bladed technical support or consult Bladed documentation for Bladed-related questions. Here are my answers to some of your questions:
- The generator mass is not specified separately in the ElastoDyn module of FAST, so, it should be lumped into the mass of the nacelle.
- Yes, the blade structural damping in the ElastoDyn module of FAST is specified in percent of critical.
- See the draft AeroDyn v15 User’s Guide and Theory Manual for detailed definitions of inputs BlCrvAC, BlSwpAC, and BlCrvAng.
- StrcTwst in the ElastoDyn module is the rotational of the principle axes of blade bending.
- Yes.
Best regards,
