I work witch OpenFast (Version: 2.1.0-183-g33a4f844-dirty)
I want to simulate the Emergency Shutdown DLC 5.1 (IEC standard). I’m using the model of the NREL 5 MW reference wind turbine. The wind turbine is seeing the conditions V_hub = 13,4 (NTM). The shutdown engages at 300s. To simulate this case I set the Parameters to:
TPitManS = 300 s
PitManRat = 8 °/s
BlPitchF = 90 °
GenTiStp = True
TimGenof = 300 s
GenDOF = True
WakeMod = 2
DBEMT_Mod = 2
UAMod = 3
In Comparison to the other forum topics
Till 320 s everything looks as expected. After that it seems the turbine is oscillation and the pitching moment gets surprisingly high. Do you have any suggestions to improve the control?
I forgot to mention, that I’m using BeamDyn instead of ElastoDyn.
Well, I’m not sure. I would guess the strong oscillation in rotor speed and blade-root moment are related. Do you get the same behavior with simplified aerodynamics, i.e., WakeMod = 1, AFAeroMod = 1?
I simulated this load case with simplified aerodynamics. The behaviour of the wind turbine looks much better, there ist no oscillation. I think there is no problem with the controll setup. So now there is the question, if a complex aerodynamics and a complex structual beam results in oscillation?
with simplified aerodynamics there is a different behaviour in the Pitch Controll. If you compare both simulations there is quite a different in the pitch angle. Does this results because of a different angle of attack?
I would first try to isolate which model is causing the strong oscillatory behavior. I would guess it is related to the unsteady airfoil aerodynamics model associated with the pitching moment, but that is just a guess. Do you get reasonable behavior with WakeMod = 2 and AFAeroMod = 1 and unreasonable behavior with WakeMod = 1 and AFAeroMod = 2?
I’m a bit surprised with your plot of pitch angle. I would normally expect dynamic wake and unsteady aerodynamics to effect the variations in response about the mean, not the mean itself. Are you getting reasonable structural response (blade and tower deflections, including blade twist) between 200-280 seconds?
I isolated both models, it seems that AFAeroMod is causing the oscillatory behavior. The structual response seems reasonable.
Here is the tower deflection. The Last ones seems a bit off.
From these results, it certainly looks like the unsteady airfoil aerodynamics model (AFAeroMod = 2) is causing the strong oscillations after the shut down. You haven’t stated how you’ve set the unsteady airfoil aerodynamic coefficients in the airfoil data files, but you may need to play with these values to get proper behavior. Reducing filtCutOff (say from “DEFAULT”, which is 20 Hz, to 5 Hz) may be a good place to start.
It looks like the switch from BEMT (WakeMod = 1) to DBEMT (WakeMod = 2) has a large effect on the mean rotor speed and blade-pitch angle between 200-280 s. This still surprises me as I would expect this change to effect oscillations about the mean rather than the mean itself. What happens from 0-200 s? What does the blade-torsion look like?
I didn’t changed the unsteady airfoil coefficients in the airfoil data files.
For example these are the data for DU30 A17:
my next step will be to reduce the frequency to 5 Hz and to look into time between 0 and 200 s.
in the the pictures are shown the the root Momentaround the x,y,z axis vom 0 to 200s with the same unsteady aerodynamic coeffizient
and here is the blade root torsion at filtCutOff = 5
Best regards Moritz
Actually, I was asking for the blade-tip twist (torsion).
And what do the rotor speed and blade-pitch angles look like from 0-200 s?
Does setting filtCutOff = 5 Hz give reasonable results during the shut down?
in the picture is shown the rotor speed ,the pitch angle and the tip torsion form 0 to 200 second with the default setting 20 Hz
Blade torsion, rotor speed and pitch angle at 0 to 400 seconds with filtCutOff = 5.
There ist no oscillation left at the results.
I’m not sure what you mean with B1N9MT (N). The tip twist would be something like B1TipRDzr (-).
The rotor speed and pitch angle look consistent (they are in Region 3 with WakeMod=1 and they are below Region 3 with WakeMod=2), but I don’t really understand why this is the case. Assuming these simulations using exactly the same wind, for some reason your results are showing WakeMod = 2 results in less blade loading than WakeMod = 1. To understand further, I would probably simplify the model (zero degrees of freedom and fixed control) and look at the aerodynamic loads (RtAeroFxh and RtAeroMzh) to see how the loads differ for this case.
I can’t tell from your plots, but do you get the response you expect with filtCutOff = 5 Hz?
I simplified the model. The wind speed is steady at 13.4 m/s. There is no degree of freedom an no pitch controll. The main difference of the aerodynamic loads is between WakeMod 1 and WakeMod 2. After pitching to feather, they are nearly the same.