Dear Dr. Jason,
I really appreciate your help and comments.
I want to get the tower first and second bending mode in the fore aft and side to side directions for IEA 15 MW on monopile.
I applied white noise by sitting
in ElastoDyn I requested the following output
the first two figures illustrates that the tower first bending mode in fore aft and side to side are 0.195 and 0.194 Hz respectively. the second tower fore aft and side to side are 1.23 and 1.4 respectively. it’s hard to get them from the first and second figures, but they are apparent in the third and forth figures.
are thses frequency right? or wrong? and how about the peaks in the red circle?
Regards,
Marwa
Dear @Marwa.Mohamed ,
I am Riad, a wind turbine enthusiast.
I look at your files.
i suggest that you put the parameter WvHiCOff in HydroDyn to 500.
In .fst file, you should disable ServoDyn.
In ElastoDyn, i prefer to set blade pitch to 90 and RotSpeed to 0.
Also, i recommend to apply the following code whether you are using MATLAB:
[pxx,f]=pwelch(tower_top_disp,[ ],[ ],2^nextpow2(length(tower_top_disp)),1/dt);
%% where dt is the time step in OpenFAST
plot(f,pxx,“r”,“Linewidth”,2)
Hope that can help.
N.B: dont forget to remove the mean value from tower top displacement time series before computing the PSD. There is a function in MATLAB called detrend
Best Regards,
Riad
Thanks Riad for your help and sharing the matlab code.
I think with the code, I could reprisent how the soil type affected both the first and second tower mode.
Regards,
Marwa
Dear @Marwa.Mohamed.,
Just for curiosity, may i ask about the simulation duration you use when you generate these plots.
Thanks in advance,
Best Regards,
Riad
I ran the models for 250 seconds not 630 seconds, will this affect the PSD shape?
Regards,
Marwa
Dear @Marwa.Mohamed,
For me, it depends.
In fact, i am working on the 5 MW barge-type FOWT. I carried out a dynamic analysis for 630 seconds. For some DoFs like surge and sway, the PSD does not show peaks associated to their natural frequency. Then, someone told me to redo the analysis for three hours and i was surprised that the shpae of the PSD for the very low frequency modes have changed.
Thanks God, it does not influence my entire work since i am interested in the pitch motion. But yeah, simulation duration does influence.
In my opinion, for bottom-fixed wind turbines, i think that 630 seconds is good.
What can you do is that for the same realization of wind and wave, you have to run two simulations: the first with 250 s and the second one with 630 s. Then, compute the corresponding PSD and compare. If they have the same shape, then 250 s is good to be adopted, if not then you should adopt 630 s.
Pay attention, the PSDs should have the same shape, no the same value of power. I mean that if the PSD of 250 sec has shown a peak at 0.18 Hz and the one of 630 sec has shown a peak at the same frequency, so this means that 250 s is sufficient to capture the dynamics of the OWT.
N.B: you should exclude the first 30 s from the 630 seconds simukation to ignore the transient effcet of the blade pitch controller. In my case (i.e. 5 MW barge-type FOWT), i ran the simukation for 11 000 seconds and i ignore the first 200 s since i see that platform surge needs almost 200 s to reach the static equilibrium position where it vibrates around.
Dont forget also to remove the mean value from the time series when you compute the PSD since if the time series has a mean value different from zero, this will induce a very high power at 0 Hz in the PSD plot.
Hope that helps.
Best Regards,
Riad
2 Likes
thanks Raed, I really appreciate your comments. I will redo the analysis considering your comments.
Regards,
Marwa
Dear Dr. Jason,
Thanks for your comments. When I further investigated the optimization results of the IEA 22 MW turbine on a monopile foundation, I found that the optimized turbines with lower natural frequencies experienced a nearly twofold increase in load compared to the original design. Although these designs did not violate the constraints related to von Mises stresses or the lateral capacity of the pile, a significant issue was observed: the blade tip deflection toward the tower exceeded the 30.0 m clearance limit. This excessive deflection may be indicative of resonance effects, as the natural frequency of the structure approaches excitation frequencies, potentially leading to amplified dynamic responses.
Could the excessive blade tip deflection beyond the 30.0 m clearance—despite satisfying stress and capacity constraints—be considered a manifestation of resonance due to the low natural frequency of the optimized structure? should I put the blade tip deflection as a constraint for the optimization?
Regards,
Marwa
Dear @Marwa.Mohamed,
We’ve discussed the natural frequencies, excitation frequencies, and resonance in past forum posts.
Regarding the blade tip-to-tower deflection, I agree that this should be included as a constraint in system optimization studies because a blade strike on the tower should always be avoided.
Best regards,
Dear Dr. Jason,
I have a question please, I hope you help me with any comments
If I have 3 turbines (NREL 5 MW, IEA 15 MW, and IEA 22 MW) installed in the same geotechnical condition, and in the same metocean data, is it normal that after the analysis of the three turbines even under fixed base monopile (no SSI) that the critical DLC differ from turbine to another? or it should be one critical DLC for all the turbines? when I analyzed the three turbines and consider PSF, the critical DLC for IEA 15 MW is DLC 5.1, for IEA 22 MW is DLC 2.3 for NREL 5 MW is DLC 6.2. Is that Ok or there is something I missed?
Regards,
Marwa
Dear @Marwa.Mohamed,
I’m not familiar with the critical load cases for all of these turbines, but it is not surprising that different turbines would have different critical load cases.
Best regards,
Thanks Dr. Jason for your comments. I really appreciate your help.
I have one more question please, which may be vey basic.
for operational DLCs, I understand the role of controller in which as the torque increases with wind speed, this causes the power to increase until it reaches the rated power at the rated wind speed. The power must be maintained at its rated power (regulated) after the rated wind speed. Since the local inflow angle is the sum of the blade pitch angle and angle of attack, a pitch-regulated blade controller will twist the blade to increase the pitch angle and lower the angle of attack. Therefore the thrust force under rated wind speed may be larger that other wind speeds between the intervals from cut in to cut out.
What I don’t understand is the parked (idling) condition, in where the wind speed is with return period of 50 years and extreme sea state and there is no controller. How this DLC 6.1 or 6.2 may be smaller than operational DLC which may have wind speed at rated wind speed?
Regards,
Marwa
Dear @Marwa.Mohamed,
I agree with your commentary about the operational controller.
For the parked/idling state under extreme wind, the controller of a modern utility-scale turbine will pitch the blades fully feathered (around 90 degrees) to minimize aerodynamic loads on the rotor when the rotor is aligned with the wind, and the yaw controller is functional to maintain this alignment as much as possible. With these settings, the mean aerodynamic load on the rotor is quite small, but the aerodynamic drag on the nacelle and tower are likely higher than that on the rotor. The standard deviation of aerodynamic loads can be still be high due to the likely high turbulence in extreme winds. Also, at sizeable yaw errors, aero-elastic instability of the idling rotor is a concern.
Best regards,
1 Like
Dear Dr. Jason,
I hope you could help me with any comments
I’m currently comparing two approaches for modeling soil–structure interaction in OpenFAST for the IEA 22 MW monopile: one using reduced stiffness matrices derived from ABAQUS substructuring and the other based on geotechnical closed-form equations. As expected, the geotechnical approach overestimates stiffness due to its linear soil assumptions—resulting in up to a 25% higher natural frequency in very soft soil compared to the nonlinear ABAQUS-based model. Abaqus stiffness approach 0.99 Hz and geotechnical approach 0.123 Hz ( very soft soil)
However, when evaluating load responses across various DLCs, I observed something unexpected. While the difference in natural frequency is substantial, the resulting load difference is relatively small (~7%) in most cases. Yet, for DLCs involving extreme sea states (e.g., DLCs 6.x, 7.1, and 1.6), the load discrepancy becomes significantly larger—up to 25–30% in some cases in parked cases and 17% in DLC 1.6.
I initially assumed such differences would appear primarily in parked DLCs due to the absence of aerodynamic damping, but DLC 1.6 is an operational case. Could you kindly advise on why the geotechnical approach’s underestimation becomes more pronounced in these ESS scenarios?
Thank you in advance for your time and insights.
Best regards,
Marwa Mohamed
Dear @Marwa.Mohamed,
The influence of system natural frequencies on the structural loads will depend on how close the excitation frequencies are to the natural frequencies and the damping of the modes being excited.
Best regards,
Dear Dr. Jason,
If I want to include the diameter as a variable in optimization for IEA 22 MW, what are the upper and lower limits for the diameter along the tower and monopile? I know that the diameter should decrease as it goes up, but is there a minimum ratio between the top tower diameter and the base tower diameter to make it reasonable? the maximum monopile diameter is 11 or 12 m, should the base tower diameter in this case be 11 or 12 m?
any comments will be appreciated
Regards,
Marwa
See this FAQ “Are the values of tower/monopile diameters realistic?” on the wiki Frequently Asked Questions (FAQ) · IEAWindSystems/IEA-22-280-RWT Wiki · GitHub
We’ve also recently seen a study reoptimizing the tower of the floating IEA22 where maximum tower diameter was 12 meters. In that case there was no limit from the drilling equipment.
Thanks for your reply.
Is there a relationship between the tower base diameter and tower top diameter, or the allowable slope to start with in the optimization.
Regards,
Marwa
Is there a ratio or limitation on the ratio between the thickness along the tower and the corresponding diameter? when I ran the optimization considering the diameter of the tower. (top tower and base tower, the intermediate is by linear interpolation). I found that some models did not run in OpenFAST and gave this error. I compared theses solutions with other solutions that run, I found that the models that have error, have high ratio between thickness and corresponding diameter.
I would appreciate any comments you can share.
Regards,
Marwa
Hi Marwa,
this tutorial can be of help 5. Tower and Monopile Example — WISDEM 2.0 documentation
The diameter to thickness ratio should usually stay between ~80 and ~200. I am not entirely sure about the tapering limits for tower diameters. You will probably need to look for those.
Note that OpenFAST does not receive in input the wall thickness of the tower, rather the stiffness and mass distributions. Although stiffness and mass clearly come from diameters and wall thickness, linking poor convergence of OpenFAST to wall thickness or ratio of outer diameter to wall thickness may become a little tricky
I hope this helps.
Pietro
1 Like
