Fatigue load reduction of bldae

Dear all,

I have a question about the reduction of blade fatigue load.

As far as I know, I can use individual blade pitch control to reduce the blade load.
In order to design the controller, I need the information of blade flap-wise root moments.
The question is, if I want to reduce the fatigue load, I need to reduce the magnitude of moments or to reduce the variation of the moments.

Thank you!

Cheng

Dear Cheng,

Fatigue loads are oscillations about a mean value. Considering the Goodman criteria, the same oscillation about a higher mean load will result in increased fatigue. Thus, ideally you’d reduce both the amplitude of oscillation and the mean, but it is likely that IPC would be most useful at reducing the amplitude.

Best regards,

Dear Jason,
Thank you for your fast reply.
I remembered I asked you a similar question about the fatigue load evaluation in this forum[url]http://forums.nrel.gov/t/question-1-for-turbulence-fields-points-outside-of-field/69/1].

I only have a single time series of results, you recommend me to use STD or RMS of moments to evaluate the fatigue load. However, just as you said, the same oscillation about a higher mean moments leads to same STD but higher RMS, I’m confused now, which performance index is better, STD or RMS.

Since I used STD to evaluate the fatigue load (of tower base and blade root) in my previous work, i.e., I want to get smaller oscillation. If I use IBP to reduce the amplitude of blade bending moments, I’m not sure I will get small STD of blade bending moments.

Best regards,
Cheng

Dear Cheng,

The “best” would be to calculate a damage-equivalent load (DEL), together with the Goodman correction to account for load means. RMS and STD are simplifications of this (with RMS better than STD).

Please note that a proper fatigue analysis would require many more than a single time series. It is standard practice to test your controller for a full range of operational conditions (from cut-in wind speed to cut-out) through many separate time-domain simulations. At the minimum, I would recommend using 2 m/s wind-speed bins and 6 turbulence seeds per bin.

Best regards,

Dear Jason.

Thank you.

Actually, no matter which index I use. I think the inputs of IBP controller have the following options, I don’t know which one is better

  1. The variation (with respect to an operating point) of the blade root moments; the control objective is to reduce such variation to zero.

  2. The rate of change of blade root moments; It is similar to the 1st case, one can reduce the rate to zero such that reduce the variation.

  3. The blade root moments; the control objective of to reduce the value of moments (same as you said, reduce the amplitude of moment). But the question is, the moments of the blade root (after the MBC transformation) are not varying around zero, I don’t know if I want to reduce them, how can I fix my reference (in the previous cases, the reference is zero).

Best regards,
Cheng

Dear Cheng,

I’m not sure I understand your question, but I would suggest reviewing papers on independent blade-pitch control to see how others have defined the objective function. I’m not an expert on wind turbine control.

Best regards,

Dear Jason,

I appreciate your help, my last question is, is there any link between the FlapDOF1 and blade root flap-wise bending moment?

Best regards,
Cheng

Dear Cheng,

I would expect a strong correlation between the displacement of the flapwise degree of freedom and the flapwise bending moment, but there will not be perfect correlation (one is not a scaling of the other).

Best regards,

Dear Jason,

I made some simulations using IBP control, I check the RMS of platform motion, DEL of tower base bending moment and DEL of blade flap-wise bending moment, but I’m not sure my simulation results are reasonable.
You can find the results in the closed file, and I made some comparisons between the three controllers GSPI and my controller (ASTW) with CBP control, and ASTW with IBP control. In Figure 2 Figure 3 and Figure 4
1 the red line is the normalized value of GSPI with CBP control;
2 the blue bar is the ASTW with CBP control;
3 the yellow bar is the ASTW with IBP control.

From Figure 4, you can find that the IBP controller reduced the blade flapwise load comparing with CBP controllers. However, the platform yaw anlge and yaw rate are increased significantly, also, the platform roll rate increased.
My question is :
1, is it normal for IBP controller have such an increase on the platform yaw and roll motion?
2, do you think the DEL of tower base side-to-side, for-aft and torsional is normal with respect to the motion of the platform?
3, concerning the mooring line force, in your point of view, is it reasonable?

You can also find the time series of all the variables in the attached file.

Thank you!

Best regards,
Cheng
res.pdf (676 KB)

Dear Cheng,

I do not have direct experience myself designing/testing IBP controllers for floating offshore wind turbines, but I’m not surprised with your results, i.e., that introducing IBP will increase yaw and roll motions. The floating wind controls research from about 10 years ago from Hazim Namik and Karl Stol comes to mind…I suggest reviewing some of their published papers, if you have not already. I would guess you’d have to introduce multiple control objectives to minimize any unintended consequences of introducing IBP.

Best regards,

Dear Jason,

Thank you for your FAST reply.
Yes, I agree with you that IBP increase the platform roll and yaw motion
However, what I am not sure of is that the tower base side-to-side DEL is decreased even the roll motion increased. Yaw motion increased significantly, but the torsional DEL increased a little bit.
As far as I know, the DEL is related to the corresponding motions, so, do you think the DEL values are reasonable?

Best regards,
Cheng

Dear Cheng,

From the plots you shared, it looks like the roll rate increased while the roll angle decreased slightly with IBP. Keep in mind that the tower side-to-side and tower-fore-aft moment orientations change with the platform motion (because the tower is cantilevered to and moves with the platform).

Best regards,

Dear Jason,

Figure 1 and Figure 5 show the conclusions as you said.
Sorry, I still don’t know if the DEL of the tower base is reasonable or not… :blush:

Best regards,
Cheng

Yes, I think the DEL is reasonable. Of course, you’ll have to run many more simulations across different wind/wave conditions and seeds to get reliable loads results.

Best regards,

Dear Jason,

Thank you.
From Figure 2. How to explain that the STD of mooring line fair-lead force and anchor force reduced with increased roll and yaw rate.
What kind of motions affects the force of mooring lines?

I didn’t find the analysis of mooring lines in the literature, but my colleague in the Hydrodymaic lab said that the mooring line loads are also important.

Best regards,
Cheng

Dear Cheng,

The mooring loads would be related to the platform motion based on the mooring configuration and line properties. I don’t know what FAST model you are using (or if you made your own), so, I can’t really comment on that.

Best regards,

Dear Jason,

I used the NRELOffshrBsline5MW_OC3Hywind model.
The _MAP.dat file is as following

---------------------- LINE DICTIONARY --------------------------------------- LineType Diam MassDenInAir EA CB CIntDamp Ca Cdn Cdt (-) (m) (kg/m) (N) (-) (Pa-s) (-) (-) (-) Material 0.09 77.7066 384.243E6 0.001 1.0E8 0.6 -1.0 0.05 ---------------------- NODE PROPERTIES --------------------------------------- Node Type X Y Z M B FX FY FZ (-) (-) (m) (m) (m) (kg) (mˆ3) (N) (N) (N) 1 fix 853.87 0 depth 0 0 # # # 2 Vessel 5.2 0 -70.0 0 0 # # # ---------------------- LINE PROPERTIES --------------------------------------- Line LineType UnstrLen NodeAnch NodeFair Flags (-) (-) (m) (-) (-) (-) 1 Material 902.2 1 2 tension_fair tension_anch ---------------------- SOLVER OPTIONS----------------------------------------- Option (-) repeat 240 120

Best regards,
Cheng

Dear Cheng,

Given that the platform rotations you are showing are small are not too different between the controllers, I would guess the difference in mooring loads are not related to the platform rotations; rather they are probably more related to surge and sway translations. Plotting those may help explain why the mooring loads differ.

By the way: MAP++ generally provides mooring reactions that enable calculation of floating wind turbine response, but mooring line loads tend to oscillate much more when using a dynamic mooring model, e.g. MoorDyn. (That is, MAP++ generally captures mean and low-frequency effects, whereas MoorDyn will capture those as well as high-frequency oscillations.) I would suggest using MoorDyn if your goal is to calculate loading of the mooring lines.

Best regards,

Dear Jason,

Thank you very much for your patience and answers, it is very helpful for me!

Best regards,
Cheng

Dear Jason,

I used individual blade pitch (IBP) control reduced the 1p load of blade root flap-wise moment, and plotted its PSD as you see in Figure 1 and Figure 2.
The simulations are made by FAST OC3Hywind model with all DOFs enabled, Figure 1 shows the PSD results of constant wind and no wave; Figure 2 shows the PSD results of stochastic wind with the irregular wave;

I found that in the literature, only the 1p load is reduced, the load under or higher than 1p frequency is almost the same as collective blade pitch (CBP) control, for example, the PSD in Figure 3, the CBP and IBP controllers have almost coincident trajectories except for the 1p frequency.

Concerning my simulation results, my question is,

1, In Figure 1, the 1p load of IBP is reduced, but the low-frequency load (lower than 1p) of IBP is also smaller than CBP, is it normal? Should it be the same as CBP control?
2, Do you think the result of Figure 2 is OK?

I’m not a specialist of load analyzing, I want to make sure my results are reliable, could you give me some comment?
Thank you in advance!

Best regards,
Cheng