I have a question about Campbelldiagram,I see the 1P , 3P,6P,12P .3P maybe the turbine have three blades.I don’t understand the 6P and 12P.
I think the 1P,2P and 3P are important.but I can’t find the 2P.Is the 2P not significative?
In the rotating frame (e.g., the blade root), you will see loads that are harmonics of the rotor speed, 1P, 2P, 3P, etc. But in the fixed frame (e.g., rotor speed, rotor torque, or tower), these loads are translated into harmonics of 3-times the rotor speed, 0P, 3P, 6P, etc., unless the rotor loads are imbalanced.
A Campbell Diagram is usually expressed with all modes converted to the fixed frame.
I don’t have a Campbell diagram for the NREL 5-MW turbine atop the OC3-monopile, but I would expect that:
The support structure natural frequencies would be practically independent of rotor rotational speed. The first bending modes of the OC3-monopile support structure are just below 0.3 Hz and the second bending modes are just below 2.5 Hz.
The rotor natural frequencies would be practically independent of the support structure. Thus, other rotor modes are likely similar to the Campbell diagram we’ve calculated for the NREL 5-MW turbine atop the OC3-Hywind spar, as found on slide 9 of the “Linearization” presentation from our most recent FAST workshop: wind.nrel.gov/public/jjonkman/Pr … onkman.pdf.[/]
Actually the Campbell Diagram shown on that slide was derived from one of my colleagues, and I don’t have the original file.
However, as mentioned in my posted dated Sep 19, 2016 in the following forum topic: AeroDyn Feedback, I’ve recently derived a Campbell Diagram of the land-based NREL 5-MW turbine with FAST v8.16 (and matched to what was derived with FAST v7). You can find the corresponding CampbellDiagram.xls file attached.
I’m presenting these results as part of a paper about the new linearization functionality of FAST v8 at TORQUE 2016, with the paper to be published soon. Here is an excerpt from the paper that describes how the Campbell Diagram was derived (starting with Test18 from the FAST v8.16 CertTest):
As mentioned in my forum post dated Sep 22, 2016 above, the FAST v8 linearization analysis used to derive the Campbell Diagram of the land-based NREL 5-MW turbine involved disabling the InflowWind and AeroDyn modules, meaning that the influence of aerodynamic loads is eliminated up front, rather than during post-processing.
Thanks for reply. Actually I was looking a case study to verify my results for NREL 5MW linearization with FAST v7. And I encounter with this topic. I tried to linearize at different rotor speed to generate the campbell diagram. But solution was not converged with RotSpeed under 9.0 rpm. (except 0 rpm because it is linearized static equilibrium point). I don’t understand what the problem is. I tried to what error message said me to do, and incrased Tmax, and changed DispTol and VelTol. It is still not solved. May you give any advice about it?
My understanding is that you are linearizing without aerodynamics i.e. CompAero = False. Without aerodynamics, my guess is using CalcStdy = False would be sufficiently accurate. What are you using for CalcStdy and TrimCase?
CompAero was set True in my analysis, I was not aware of CompAero should be False to exclude aerodynamics. In fact, that was what I try to ask in my first post (Nov 18, 2016). I probably couldn’t make myself clear.
Also, I use CalcStdy=True and TrimCase=2, because I was not sure about the initial conditions in primary input file. So I just wanted to let FAST calculate the operating point. Can I get the accurate solution if I only change the RotSpeed when CalcStdy=False (while all other initial conditions are set zero) ?
One last question. So, should I set CalcStdy=False while CompAero=False?
You can use the linearization options that you want, but when aerodynamics are disabled (CompAero = False), the trim solutions may not function properly. FAST v8 allows you to linearize at any point in time, without a trim solution, so, it is probably easier to use for cases without aerodynamics. When aerodynamics are disabled, the only other load is gravity, and, depending on the case, it may be sufficiently accurate to linearize about the initial conditions.
I try to identify the natural frequencies of an offshore FAST model. so i linearized the NREL 5MW Onshore Wind Turbine (from Test18) with FAST v7 to get a Campbell Diagram. But i don’t know if my data for blade “collective”, “progressive” and “regressive” are exact . furthermore my Drivetrain frequency behaves like for OC3 Hywind Spar.I supose my input file is for OC3 Hywind Spar. but i can’t found any file comparable to Test 18 to linearize with FAST 7. Can anyone help me please. Here is an is my Result.
From my quick review, your Campbell diagram looks reasonable to me. The drivetrain frequency looks odd at 0 rpm because in FAST v7, FAST would disable the generator DOF automatically when linearizing with the generator speed set to zero. The linearization functionality of FAST v8 does not have this limitation (if you choose to upgrade). You can get around this in FAST v7 linearization by choosing an operating point that is very close to zero speed (instead of exactly zero).
I’m not sure what you mean be Test18 in FAST v7, as the FAST v7 CertTest was limited to 17 cases. If you are asking for a land-based model of the NREL 5-MW turbine that is compatible with FAST v7, you can find that here: wind.nrel.gov/public/jjonkman/NR … nshore.zip.
Thank you so much for the clarification. I tried it with 0.1 rpm and the result is well. Now i understand why the Generator DOF was not in the Linearized data by 0 rpm. Also thans for the land-based model of the NREL 5-MW turbine. That is eyactly what i needed.
now I linearized the FAST v8 to get the Campbeldiagram like with FAST v7. But the natural frequency of all DOFs are not changing with rotor speed. including the blade. I don’t know what is wrong in my linearization. i have attached the new Cambell diagram.
I noticed a couple problems from my quick look at your input files:
In the ServoDyn input file, you’ve enabled the simple variable-speed controller, but without wind, this will cause the rotor to slow down i.e. your steady state solution is likely not the rotor speed you set. Instead, you can disable the generator from producing torque altogether by setting TimGenOn > TMax.
In the FAST primary input file, you should linearize at equally spaced azimuth steps. For 12 rpm, this would be linearizing once every 0.138888 seconds for 10-degree steps.
I have somee questions about natural frequencys. i want to identify that in nonlinear model with FFT and compare with my result off Campbell diagram. But it is dificult to find out frequency. for the exitation i found some thing about “with white noise (e.g., broadband turbulence)” in your older posts. my question is, can i add the noise on the control input like Torq control oder there is a file for broadband turbulence in FAST? also i don’t know if i can found all frequency on each output of FAST. her you can find a data with a noise example.