Dear Jason,
Thank you for your help. Turns out that the turbine was simulated with a DLC 6.1.
I although have another question ; I am noticing that in the scenarios with the Normal Turbulence Model, the user cannot decide the Wind Turbine Class , and Turbsim sum file does not provide any evidence for this parameter whatsoever.
Why is that? Wouldn’t the tubine class had a significant role in our simulations?
Thanks a lot in advance.
Dear @Nikos.Christodoulaki,
The wind turbine class in IEC influences the specification of the turbulent extreme wind model (EWM). The wind turbine class (1,2,3 in TurbSim representing I,II,III in the IEC) is specified when the EWM model is selected via TurbSim input IEC_WindType, e.g., setting IEC_WindType = “2EWM50” will result in the 50-year EWM for wind turbine class II.
Best regards,
Hi Mr Jonkman
I saw a thesis of yours under the title “Dynamics Modeling and Loads.”
Analysis of an Offshore
“Floating Wind Turbine”
Somewhere you talked about pitch fault simulation, which I have highlighted in the image below.
My question is, did you model this case through Simulink or through the primary input file of FAST software
(If possible, explain to me in detail how you did it both ways)
Dear @Ali.Rouhbakhsh,
These fault cases were run by changing the FAST input files (not through Simulink). These fault cases can be simulated by setting the override pitch maneuvers (ServoDyn inputs TPitMan, PitManRat, BlPitchF) and turning off the generator (ServoDyn inputs GenTiStp, TimGenOf).
Best regards,
Why should the generator be turned off?
Nice to connect. I am also trying to run the DLC 2.2
Can you help me what faults you modelled for DLC2.2?
Best regards,
Karthik
Dear @Karthik.Prakash,
I have not ran DLC 2.2 myself before, but here is what IEC 61400-1 Ed. 4 says about this case:
For DLC 2.2, control system failure events or internal electrical and mechanical system faults
with expected failure mode return period greater than 50 years shall be considered as
abnormal.
For DLC 2.2, rare events that have relevance for the wind turbine loading, including faults
relating to activation of secondary layer protection functions, shall be considered as
abnormal. Such faults may include erroneous activation of actuators, non-activation of braking
systems, and blocking of the pitch system. This load case shall at least address the following:
independent overspeed protection, generator overload/fault protection, uncontrolled blade
pitch protection (blade pitch runaway), uncontrolled yaw protection and excessive vibration or
shock protection.
Best regards,
Dear @Jason.Jonkman
I am trying to imitate the result of the load analysis in the IEA 15MW UMaine VolturnUS-S Technical Report.
(https://www.nrel.gov/docs/fy20osti/76773.pdf, pages 34–39)
I know there must be a difference between the IEA result and mine because the model and parameters couldn’t be exactly the same, but the difference is huge, especially in DLC 6.3.
Also, there is too much vibration in some time series data.
But I don’t know exactly what the cause is, even though I checked all the relevant posts in this forum.
Would you please check my DLC 6.3 OpenFAST parameter?
FYI, OpenFAST: v2.6.0, IEA 15MW Model: v1.1.2, ROSCO: v2.5.0
Result: IEA vs JNU (my result)
1-1) Nacelle Inertial Acceleration Maximum Magnitude Values
1-2) Nacelle Inertial Acceleration Time Series Data
2-1) Deflection and Displacement Maximum Magnitude Values
2-2) Deflection and Displacement Time Series Data
3-1) Moment Maximum Magnitude Values
3-2) Moment Time Series Data
4-1) Force Maximum Magnitude Values
4-2) Force Time Series Data
OpenFAST Parameter (e.g., 6p3_EWM1_Yaw+20_V38_S1)
Fst
ElastoDyn
Inflow
AeroDynv15
ServoDyn
DISON.IN
HydroDyn
MoorDyn
Best regards,
Sangwon
Hello Sangwon,
I’ve asked my colleague Mayank to help here. Mayank has recently been looking at DLC 6.x for the IEA15 and he’s run into many similar problems. One trick that sometimes works is to pick one of the three blades and pitch it to 70 or 80 deg (and not 90). That often reduces the vibrations, which are caused by angles of attack deep in the negative range. Mayank has also removed the 3d corrections in the lift curves of the airfoils, although he had to run our framework WEIS to do that. Lastly, I would not rely too much on the loads from the report in pdf. The OpenFAST model has undergone too many changes at this point.
I hope this helps.
Regards,
Pietro
1 Like
Dear @Pietro.Bortolotti
I really appreciate your kind and detailed reply.
Please also send my gratitude to Mr. Mayank.
Have a good one!
Best regards,
Sangwon
Hi @Sangwon.Lee,
Sorry for the delayed response.
As @Pietro.Bortolotti mentioned, I have run DLC 6.x for the IEA15 and come across similar issues that you are facing.
A few things for you to look into:
- Start with a simpler model: I would start by removing HydroDyn (CompHydro=0) and MoorDyn (CompMooring=0) and ServoDyn isn’t needed so (CompServo=0).
When setting CompHydro=0 do not forget to turn off the platform DoFs to False (Ptfm**DOF) in the ElastoDyn file.
You can add back the platform and mooring once you have the model working without them.
-
Using BeamDyn for the blade (CompElast = 2): This change will significantly increase your computation time given the smaller time steps needed. But I have observed it to improve the model for the high yaw angle cases.
-
Using Unsteady airfoil aerodynamics: I have also noticed that activating unsteady aero improved the model (AFAeroMod = 2, UAMod=3). But this comes with a lot of caviates, so would be the last thing you might want to try.
-
Removing 3D corrections: I do not expect this to improve the model stability but is for correctness.
-
Update OpenFAST: OpenFAST is currently on version 3.5.1, so I would recommend changing over if possible.
I hope this helps
Regards,
Mayank
1 Like
Just two small comments in case they are useful:
- Regarding the use of unsteady airfoil aerodynamics, I guess that during this simulation you are in deep dynamic stall conditions. My understanding is that the conventional unsteady aerodynamic theory is not valid. Maybe I’m missing something? So, personally, I would leave AFAeroMod = 1.
- Looking at the ElastoDyn input file, it seems that the generator is locked (GenDOF = False, RotSpeed = 0). This effectively means that the rotor is locked in place (i.e., parked condition). Is this the expected behavior or you wanted to leave the rotor in idling conditions (GenDOF = True)?
1 Like
Thanks for catching that Roger!
Yes, by default OpenFAST switches over to steady aerodynamics at AoA >45-deg; the aerodynamic damping does help with the model stability. But I agree with keeping AFAeroMod = 1 if you do not tune the UA model.
2 Likes
Dear @Mayank.Chetan
I sincerely appreciate your detailed description!
As you mentioned, I will try them one by one.
Thanks again, and have a good one!
Best regards,
Sangwon
Dear @Roger.Bergua
First, I appreciate your advice for my trouble.
I set ‘GenDOF = False’ to make the rotor lock.
Best regards,
Sangwon
Hi @Jason.Jonkman,
I have a basic doubt regarding simulating the extreme events. What exactly is the difference between a wind turbine in parked and idle condition in the physical sense? I have simulated both the cases, but there is not much appreciable changes in the platform responses. So which wind turbine condition is recommended to be used in an extreme event?
I am simulating a steady wind field of 49m/s.
Please clarify my doubt.
Thanks in advance.
Dear @Arya.Thomas,
“Parked” means that a brake is applied and the rotor will not rotate. “Idle” means that no brake is applied and a blade with blades feathered may slowly rotate or dither around a bit as a result of residual aerodynamic torque. In modern utility scale wind turbines, it is much more common to idle a rotor in extreme storm conditions (the brake would only be applied for maintenance purposes).
Best regards,
Thanks for the clarification 