Mooring Line

Hi Jason:
I am trying to get a better understanding of sea state that is in inputted to HydroDyn. I used a JONSWAP spectrum with gamma factor = 2.27, Hs =0.73 m and Tp = 7.44 s (fp = 0.134 Hz); 100 m deep water; Integration time step, DT, in .fst file is 0.025. I ran simulation and used the wave elevation time series to get the wave spectrum; using a sampling frequency of 1/DT (80 Hz). However, for wave spectrum created, the peak frequency is not 0.134 Hz as I was expecting. I got a peak frequency of 0.1475 Hz (Tp= 6.77 and not 7.44 s). If I play around with sampling frequency I can get a fp of 0.134 Hz. I am wondering if something is wrong with original sampling frequency I used or other error. I was expecting to get fp = 0.134 Hz (Tp=7.44 s) which was entered in HydroDyn on generation of wave spectrum. I would be very grateful for any clarification/insight that you may have about this.

Thank you.

Regards,
AOAW

Dear Andre,

A few clarifying questions:

  • You say you are using an integration time step (DT) of 0.025 s and a sampling rate of 1/DT = 80 Hz, but 1/0.025 s = 40 Hz. Which numbers are correct?
  • What do you mean that you “play[ed] around with the sampling frequency”; do you mean your changing DT_Out?
  • What have you set WaveDT and WaveTMax to? The wave elevation time series is created by HydroDyn with these values, but the time-series output from HydroDyn will be interpolated within this time series at a rate of DT_Out. This interpolation will have some effect on the spectrum.
  • Have you set WaveNDAmp = True or False? Enabling WaveNDAmp will normally distribute the wave amplitudes so the spectrum would not follow the JONSWAP spectrum exactly.

Best regards,

Jason:
See clarifications below.
-You say you are using an integration time step (DT) of 0.025 s and a sampling rate of 1/DT = 80 Hz, but 1/0.025 s = 40 Hz. Which numbers are correct? 40 is correct, I made an error.

  • What do you mean that you “play[ed] around with the sampling frequency”; do you mean your changing DT_Out? No. I changed the sampling frequency in my script that I used to convert time series from OpenFAST output to spectrum.
  • What have you set WaveDT and WaveTMax to? The wave elevation time series is created by HydroDyn with these values, but the time-series output from HydroDyn will be interpolated within this time series at a rate of DT_Out. This interpolation will have some effect on the spectrum. I used 0.25 for WaveDT and 4600 s for WaveTMax; simulation time was 1800 s. DT_Out was 0.025 s.
  • Have you set WaveNDAmp = True or False? Enabling WaveNDAmp will normally distribute the wave amplitudes so the spectrum would not follow the JONSWAP spectrum exactly. WaveNDamp has been set toTrue. Should it have been set to False?

Let me know if this now helps with unraveling the issue.

Thank you.

Regards,
AOAW

Dear Andre,

I’m still not sure what you mean by “sampling frequency”, but I suspect this is what is shift in peak frequency. With WaveNDAmp = True, the spectrum of the time series generated by OpenFAST will be quite noisy–the amplitudes normally distributed around the target JONSWAP spectrum. So, resampling could pick different peaks and valleys. Using WaveNDAmp = False will ensure that the amplitudes at each frequency follow the JONSWAP spectrum exactly, where I would expect resampling to have less an effect.

Best regards,

Thanks Jason.
I will set WaveNDAmp = False and see how that affects the results. The sampling frequency is one of the variables to get the spectrum when using Welch’s or pmtm method in MATLAB.

Regards,
AOAW

Hi Jason:
On Oct 18, 2021 you wrote:
“In the absence of wind and wave excitation, then the heave motion will remain stationary if the vertical buoyancy balances with the vertical mooring pretension and full-system weight of the FOWT (floater + tower + RNA).”

Regarding the mooring pretension:

  1. Was this value determined by wind turbine manufacturer and is it given in one of the OpenFAST input/output files?
    1a) If not given by manufacturer how is pretension determined?
  2. In which module would I be able to locate it, I checked HydroDyn and MoorDyn but did not see a value for it?

I have a few more queries:
3) Are the masses of the Floater, tower and RNA in ElastoDyn? I see hub, nacelle and platform masses. Are the mass of the blades and the rotor included in the hub mass?
3a) Is the mass of the tower accounted for in ElasToDyn_tower file?
4) I see there is platform inertia in roll, yaw and pitch. I was wondering why none included for the 3 horizontal DOFs? Just want some clarity here for further understanding.

Thank you.

Regards,
AOAW

Dear Andre,

Regarding (1,2), the vertical mooring pretension is not really a standard output from OpenFAST or the mooring modules. But you can infer it by, e.g., outputting the tension along each mooring line at the fairleads, computing the vertical contribution (based on the inclination angle of the line), and summing. You could also get it by reviewing operating point values from an OpenFAST linearization analysis. Regardless, I’m not sure you need to know the value exactly so much as to set up the OpenFAST model such that it does not heave in the absence of wind and wave excitation.

Regarding your other questions:
3) Yes. The blade and tower masses are defined (as distributed per unit length along the blade and tower) in the respective blade and tower files of ElastoDyn. If BeamDyn is enabled, then the blade mass is accounted for in BeamDyn instead of ElastoDyn.
4) In translation, the inertia is the platform mass. The platform mass and center of mass offsets also couple the translation and rotation motions.

Best regards,

Thanks Jason.

Regards,
AOAW

Hi Jason:
I have two queries about response amplitude operator (RAO) and the power spectrum:

  1. How are the RAOs computed in OpenFAST? I saw a paper in which they seemed to have used OpenFAST to compute the RAOs
  2. I used a wave only condition and varied the sea states by changing WavePkShp only. For one option I used WavePkShp = 1.00 (PM spectrum) and for the other option I used WavePkShp = 3.1(JONSWAP). Then I created power spectrum of the heave motion time series. I was expecting the energy to be greater for WavePkShp=3.1 but I realize that only occurs at Tp/fp. For lower values than fp, WavePkShp=1.0 has more energy. See attachment. Can you please provide some clarification as I had thought a higher Gamma factor meant greater energy throughout for particular a spectrum?

Thank you.

Regards,
AOAW

Dear Andre,

I would recommending computing RAOs from OpenFAST using a white noise spectrum (WaveMod = 3), as discussed in our ISOPE 2013 paper: nrel.gov/docs/fy13osti/58098.pdf. A MATLAB script (RAO.m) for calculating RAOs based on the approach outlined in this paper is available in my post dated Mar 09, 2018 in the following forum topic: RAO OC3Hywind Turbine.

Increasing the Gamma factor (WavePkShp) does not increase the energy throughout the entire spectrum; rather, increasing the Gamma factor concentrates the energy at the peak-spectral wave period/frequency, reducing the energy at other frequencies. See, e.g., a plot here: researchgate.net/figure/Spe … _308804645.

Best regards,

Many thanks for that clarification Jason.
You said the following,“… increasing the Gamma factor concentrates the energy at the peak-spectral wave period/frequency, reducing the energy at other frequencies”.

Does this mean that with increased Gamma there is less possibility for resonance at the lower frequencies? Trying to understand what this means in relation to the floating structure.

Thank you.

Regards,

Dear Andre,

Well, most floating structures are designed so that there natural frequencies are outside the range of significant first-order wave energy. So, resonance of floating structures is typically dominated by second-order (slow-drift, sum-frequency) effects rather than by the direct first-order wave spectrum. Of course, second-order effects exist from nonlinear interactions of first-order terms, so, there is an indirect influence.

Best regards,

Thanks Jason. Noted.

Regards,
AOAW