Dear @Jason.Jonkman, you were right; it was my bad not updating the Inflowind submodule with the new information obtained in the fresh wind file created by TurbSim. Now everything seems to be okay. By the way, this TurbSim manual was of great help; thank you very much.
On the other hand, in the case of the help wanted, I don’t know if I can provide any assistance in terms of testing, debugging or compiling; then, I am new at OpenFAST. But you can count on me if there’s any need for transcriptions or creating documentation about what I’m undertaking. I will check this link you gave me and see if there is any way I can cooperate.
Kind regards,
Emanuel M. Rergis
1 Like
Dear Dr. @Jason.Jonkman:
Now that I adjusted all the parameters to run some simulations, I have found an interpretation problem. I need to visualise only the tower tip fore-aft and side-to-side displacements, but I’m confused since it is still unclear to me which output parameter I must plot. Are these TwrTpTDxi and TwrTpTDyi? or TwHt1TPxi and TwHt1Py? Could you please clarify that for me? Maybe, it’s another output parameter; I’m not sure. I checked the output parameters list many times, but it is a bit unclear to me.
And one more question: why must I change the grid when creating a new wind file to run a 600s simulation from TurbSim? Is that because the original grid is insufficient to cover the whole nacelle and blades during 600 sec?
Thank you very much in advance.
Best regards.
Emanuel M. Rergis
Dear @Emanuel.Rergis,
ElastoDyn outputs TwrTpTDxi and TwrTpTDyi will provide the fore-aft and side-to-side displacements of the tower top including both the tower elastic deflection and substructure motion. These sound like the outputs you want.
ElastoDyn outputs TwHt1TPxi and TwHt1Pyi are similar except that they are the deflection outputs at the 1st tower gage node defined by ElastoDyn input TwrGagNd (which cannot be located exactly at the tower top).
Regarding TurbSim, the wind file you were using was too short to run a 600-s simulation. Regardless, it is typical to run multiple 600-s simulations (each with a different seed in TurbSim) at each condition (wind speed, profile, turbulence level) that you want to simulate, so, running TurbSim many times is likely required anyway.
Best regards,
1 Like
Dear Dr @Jason.Jonkman:
Thank you one more time for your reply. The hint you provided is very useful. Now, I’m trying to estimate the blade passing frequency. For this matter, I considered using the Azimuth angle converted in radians and dividing each value of the corresponding rotor speed by these azimuth angle values. Does it make any sense to you? I have a hunch that this is not entirely correct. If my guess is right, what have I found? Please, find attached the corresponding figure and tell me, if possible, your point of view.
Kind regards!
Emanuel M. Rergis
Dear @Emanuel.Rergis,
The blade-passing frequency (i.e., the rate at which a blade passes by the tower) for a 3-bladed rotor should be 3 times the rotor speed expressed in revolutions per second. So, output the rotor speed, convert to Hz, and multiply by 3.
Best regards,
Thank you again for your feedback and invaluable information, Dr @Jason.Jonkman. After taking all the information you provided, I started plotting the tower tip fore-aft displacements. The results of the plot do not quite match what I expected.
First, I expected the displacements to oscillate around zero, but in the plot, they do it around 0.3 m approximately. I remember that the only different initial condition I set was 1 m displacement in the fore-aft direction.
I must highlight that I’m using only one TMD; the other two are supposed to be off. I’m using the TMD to act throughout the x-direction (TMDx). The k and c values are, respectively, 64,000 N/m and 10,000 N/m/s. The mass of my TMD is 20,000 kg. The wind load is a von Karman spectrum with 15 % turbulence and a mean wind speed value of 18 m/s. My wind shear power law exponent is 0.2. Could you clarify, if possible, what could have gone wrong?
My second question would be, I also expected “more fun” in the behaviour of the baseline displacements compared to the behaviour of the displacements subdued to a TMD with a 13 % damping factor. Why am I obtaining a similar response? How could I get a more significant response from my TMD? Please, find attached a snapshot of this simulation experiment.
Thank you very much, and best wishes!
Emanuel M. Rergis
Dear @Emanuel.Rergis,
The mean value of tower displacement likely arises from a combination of rotor thrust and rotor-nacelle assembly weight overhanging-weight induced tower deflection.
It looks like the TMD is having some effect, just not as much as you expect. It has been a while since I’ve studied TMD design. Can you clarify how you chose the TMD mass, stiffness and damping and what motion reduction you are expecting?
Best regards,
Dear @Jason.Jonkman :
As I stated in previous messages, I’m trying to replicate M.A. Lackner’s results. In this specific case, he assumes the natural frequency of the tower’s first bending mode to be 1.79 rad/s. Therefore he chose a TMD mass equal to around 8% of the nacelle mass. Similarly, the values of c and k are determined based on this criterion (64,000 N/m and 10,000 N/m/s). Maybe, if I increase the mean wind speed value, I can see more vibration.
Going back to the tower’s first bending moment mode, in M.A. Lackner’s paper, this value is assumed to be 1.79 rad/s, but it clearly states differently in the NREL 5MW Baseline User Guide. Should I consider retuning my TMD in accordance with the value indicated in the user’s guide?
Please, find attached the corresponding snapshot.
Kind regards,
Emanuel M. Rergis
Dear @Emanuel.Rergis,
The table you show provides natural frequencies for the land-based version of the NREL 5-MW baseline turbine. Is this the version you are simulating? Different offshore support structures will have different natural frequencies.
Best regards,
Dear Dr @Jason.Jonkman :
Thank you for your response. I’m trying to simulate the NREL-5MW_OC3Mnpl wind turbine. I firmly believe that this is not a land-based 5-MW wind turbine. Does it mean I should use the natural frequencies reported by M.A. Lackner et al. because they use an NREL-5MW monopile wind turbine?
Best wishes,
Emanuel
Dear @Emanuel.Rergis,
The NREL 5-MW baseline wind turbine atop the monopile in 20-m water depth with fixed foundation has a 1st support structural natural frequency around 0.275 Hz = 1.73 rad/s, as given in Figure 6 of the IEA Wind Task 23 OC3 report: https://www.nrel.gov/docs/fy11osti/48191.pdf.
Best regards,
1 Like
Dear Dr @Jason.Jonkman:
Thank you once again for your reply. The manual you provided has been immensely helpful. Now I have been trying to run some tests with different wind spectra. Unfortunately, I found a problem in my simulations when setting the Uref parameter to 27.7 m/s (100 km/h), then the command prompt displays a fatal error message and aborts the simulation process. Is there anything that I am not considering here? Please find attached some snapshots with the details.
Thank you very much.
Best wishes,
Emanuel M. Rergis
Dear @Emanuel.Rergis,
What fatal error message you are receiving?
Best regards,
Dear Dr @Jason.Jonkman, please find attached the corresponding snapshot.
Kind regards,
Emanuel
Dear @Emanuel.Rergis,
It looks like your model has gone unstable.
My understanding is that your model runs well at some wind speeds, but for other wind speeds it goes unstable. This suggests to me that you are not setting proper initial conditions, particularly for rotor speed and blade pitch (resulting in the controller not be able to recover). In general, you should always initialize an OpenFAST model with rotor speed and blade-pitch set to the expected (mean) value appropriate to the mean wind speed you are simulating. For the NREL 5-MW baseline wind turbine, see Figure 9-1 in the specifications report for information on what rotor speed and blade-pitch you should be using at each wind speed.
Best regards,
Dear Dr @Jason.Jonkman:
Thank you for your response. The problem got fixed by using the Figure you advised. For the moment, I went back to basics. I removed the original 1 m displacement in the fore-aft direction and used the frequency of 0.275 Hz = 1.73 rad/s to tune my TMDx. Unfortunately, I still don’t get more contributions from the TMDx. Could this be something that I am not considering? I’ll give you the values I used:
ω=1.73 rad/sec,
k= 36,000 N/m,
c= 10380 N/(m/s)
m= 12 000 kg.
M_nac = 240 000 Kg
Uref=18.2 m/s
Kaimal Spectrum.
Alternatively, I think I should excite the tower and the nacelle so that vibration without the TMDx will be more violent. For this purpose, I plan to increase the value of Uref to approximately 100 km/h (27.78 m/s). Is there any other way to obtain more aggressive displacements? My aim here is to prove the TMD theory right. What can you suggest to me? Please find attached the corresponding snapshots.
One more thing, if I removed the 1 m initial displacement in the fore-aft direction, why do you think I still have an initial displacement that goes up to around 1.8 m?
Thank you very much, and kind regards!
Emanuel M. Rergis
Dear Dr @Jason.Jonkman:
Thanks once again for your feedback. Regarding the effect of the TMD, I can assume that more frequencies are affecting my system. That’s the conclusion I can reach when I look at the FFT I applied to the Tower tip displacements (please find attached the snapshot). In this plot, I found very low frequencies having influence. Are those the ones that I am supposed to tackle as well?
One more thing: I must start implementing a Pendulum Tuned Mass Damper into this OpenFAST model I have been testing. How can I start this? I already have my equations of the PTMD and the electrorheological damper attached to the nacelle ground. How can I make MatLab appropriately interact with OpenFAST? Should I change the source code? How could I add more options to the ServoDyn module?
Thank you very much, and kind regards!
Dear @Emanuel.Rergis,
I can’t say how you should design your TMD, but I would assume you’d want to add damping to the natural frequency(ies) and/or excitation frequencies (1P, 3P, 6P).
OpenFAST does not currently support augmentation of the TMD model through MATLAB/Simulink. So, if you can’t represent your “pendulum tuned mass damper” with the existing StC submodel options of ServoDyn (TMDs), it sounds like you’ll need to change the source code. My guess is all of your changes could be isolated within the StC source code. If necessary, I would suggest adding support for a new StC_DOF_MODE. I suspect the approach would be similar to that for the TMD models. So, as a first step, I would ensure that you have a thorough understanding of how the TMD model is implemented in StC, e.g., by ensuring that you understand the theoretical basis of TMD (4.2.12.2. Theory Manual for the Tuned Mass Damper Module in OpenFAST — OpenFAST v3.4.1 documentation) and that you understand how this theory has been implemented within the StC sourc code. After that, I would suggest writing down the theoretical basis of your “pendulum tuned mass damper” in a form that mimics the TMD basis. Source code modifications would hopefully be straightforward after that.
Best regards,
Dear Dr @Jason.Jonkman:
Thank you one more time for your responses. After checking all the information about the NREL 5-MW baseline wind turbine atop the monopile in 20-m water depth with a fixed foundation and revising the documentation you provided (thank you once again), I think I might need to change the StC source code. Nevertheless, before I start getting familiar with the source code, I need to know some information that is unclear to me in the documentation. Could you clarify the height of the nacelle in this model? I don’t manage to find that data.
Kind regards,
Emanuel M. Rergis
