Dear Dr. Jason,
thanks for your comment, I could decrease the time successfully. Now, I hade a new problem.
I output the mode shape from SAP and then import the coefficients of 6 order polynomial instead of BModes.
whe I tried to run the model
which means the model is unstable, when I further investigate, I found that the problem when I call the mode shape from SAP, their direction changes with changing the tower geometry and/ or base condition. what I mean with direction change is mode#2 for example is SAP in some models will be 1st mode in y direction and other cases it will be 1st mode in x direction. some times the mode #3 is tortion.
Is there any way to solve this issue? Do the mode shapes direction (x or y) differ with the tower thicknesses or base condition (flexible or fixed)
Regards,
Marwa
Dear @Marwa.Mohamed,
The first-tower fore-aft and side-side modes often have very similar natural frequencies, and perhaps your small changes to the substructure or tower design cause their order (from lowest to highest frequency) to switch. I would simply ensure that the fore-aft and side-side modes are properly selected from the SAP output based on which modes have x- and y- deflection.
Best regards,
1 Like
Thanks Dr. Jason for your quick reply
Dear Dr. Jason,
I had another problem, it would be appreciated if you help me with any comments.
In my problem, the optimization suggests any thickness for the tower between 0.004 m and 0.2 m based on paper “Conceptual monopile and tower sizing for the IEA
Wind Task 37 Borssele reference wind farm”. I solved the problem of coefficients of 6 order polynomial for the mode shapes. But the problem that I have this error message
when I investigate the ElastoTower , I divided the tower to 9 segments, as shown in this figure
the thickness of the tower base =0.189 m, while the remaining thicknesses = 4.58e-3 m. I know that this thickness is very small, but this thickness was within the limits. the only condition I put that the upper thickness is less than the lower thickness (to have tapered, and avoid the lower sections to be smaller than upper section). the optimization dive a solution of t2=t3…=t9=4.58e-3. Is this could be the reason for the error? or not?
when I tried to decrease the time step from 0.01 to 0.05 sec. I had this error message
Thanks for your time
Regards,
Marwa
some trials , I had this error also. what is that mean?
Regards,
Marwa
Dear @Marwa.Mohamed,
Well, I see several warnings before the fatal error regarding large blade and platform deflections, low tip-speed ratio (TSR), BEM not finding a solution (“no valid value of phi”), and large inflow-skew angle (chi). This tells me that your model is unstable. I would guess the large deflections are triggering the aerodynamic warnings and eventual error. Presumably the large deflections are driven by the changes to your tower properties and mode shapes in ElastoDyn or your SubDyn model. But it is hard for me to know exactly what based on the information provided. As with any such issue, I would review your inputs for correctness and simplify the model set-up to debug, e.g., isolate the issue to a specific problematic tower or platform degree of freedom.
Best regards,
1 Like
Thanks Dr. Jason for your detailed reply.
The main idea that the optimization suggests thicknesses, then theses thicknesses and mode shapes of the tower FROM sap sent to OpenFast to generate openFast model, then the updated load from Openfast sent to SAP model with (considering theses updated load from the OpenFAST after considering the updated thicknesses and mode shapes), then apply the constrains (frequancy, stresses, deflection…).
when I further investigate the models of the solutions that have error. I applied white noise as shown in theses figures. then I investigate the frequences of theses models compared to other models that run successfully.
I found that the models that have warrnings and message errors I sent before, has natural frequences around 0.08-0.09 which are close to the wave excitaion frequency and that lead to resonance, that’s why the model can not run. Is my conclusion is correct or not please?
Regards,
Marwa
Dear @Marwa.Mohamed,
Your explanation sounds plausible, but your PSDs are a bit hard to interpret because the frequency resolution is quite coarse (likely due to too short of a simulation). If your first support-structure-bending natural frequencies are indeed in the 0.08-0.09-Hz range, these sound too low to be realistic.
Instead of white-noise excitation, you could verify the full-system natural frequencies through a linearization analysis followed by Eigenanalysis, e.g., using the Automatic Campbell Diagram Code (ACDC) to assist: GitHub - OpenFAST/acdc: ACDC: Automated Campbell Diagram Code.
Best regards,
1 Like
Thanks Dr. Jason for your help and time.
Regards,
Marwa
Does my implementation for white noise excitation is correct? or should I put CompAero=0 as well ?
Regards,
Marwa
Dear @Marwa.Mohamed,
I see that you enabled white-noise wave-excitation in SeaState and disabled InflowWind. Keeping AeroDyn enabled without InflowWind will mean that AeroDyn is making use of still air, which is fine if you’ve disabled the wake and unsteady airfoil aerodynamics models of AeroDyn.
Best regards,
1 Like
Dear Dr. Jason,
I have a question please, how the floating OWT is designed in term of natural frequency? for monopile OWT it has to be is soft- stiff zone between 1P and 3P and away from wave excitation frequency. but how about floating, according to paper “Geotechnical challenges in monopile foundations and performance assessment of current design methodologies” by S. Jindal et al., floating it should be in the soft soft zone. So, how about the wave excitation frequency which is very close to the 1st tower natural frequency
I asked this because I compare the stiff and soft soil for monopile OWT in term of the optimization, the load applied incase of soft soil is bigger than the case of stiff soil. and the optimzation resulted in a solution with natural frequency of (0.146 for stiff compared to 0.134 for soft) which is very close for wave excitation frequency 0.13.
Regards,
Marwa
Dear @Marwa.Mohamed,
Typically the rigid-body natural frequencies of a floating OWT (surge, sway, heave, roll, pitch, yaw) are placed either just below or just above the first-order wave-energy range, depending on the floater type. Typically the first tower-bending natural frequencies–which are higher in floating systems than in an equivalent fixed system due to the free-free boundary condition–are placed in the soft-stiff region, although they could be in the stiff-stiff region for large turbines (around 15-MW and above).
Best regards,
Thanks Dr. Jason for your reply. So in case of IEA 22 MW, which has 1P rated+5% = 0.1236 Hz and 3P rated-5% = 0.335 Hz. and the wave excitation frequency in U.S.S East Coast = 0.1-0.13 Hz.
this means that my models from optimization should be higher than 0.13 Hz or should it be further far away like minimum 0.15 Hz?
Regards,
Marwa
Dear @Marwa.Mohamed,
Presumably you are referring to the first tower-bending natural frequencies? I would ensure this frequencies are not within the linear wave-excitation range, but have some margin above. Also at 22-MW, the tower may need to be stiff-stiff, not soft-stiff.
Best regards,
Thanks Dr. Jason for your quick reply. I want to make sure I understand your comment correctly please
1- If the turbine will be installed in U. S. East coast which hast peak spectral period ranging between (9.452 sec to 7.441 sec), in that case, the wave excitation range is between (0.106-0.1344) so the margin for the first tower bending natural frequency should be above for example 0.15?
2- if the turbine will be installed in Gulf of Mexico, which has peak spectral perid ranging between (9 and 5.707), this means that wave excitation frequencies between (0.11 to 0.175) so the margine will be at least for example 0.2 ?
3- from the manual of IEA 22 MW they stated that
Does that mean that the turbine is desgined based on soft - stiff ?
4- and the first tower natural frequency is calculated from BModes which will be categorized as soft-stiff or stiff-stiff, but the full turbine system can be calculated from white noise or free decay or OpenFAST linearization, right?
Regards,
Marwa
Dear @Marwa.Mohamed,
Regarding (1-3), I agree with your comments. In my previous post, I was referring to a floater support structure for the IEA Wind 22-MW RWT. I agree that for the fixed-bottom monopile-based support structure, the tower is designed as soft-stiff.
Regarding (4), OpenFAST can certainly be used to compute the full-system natural frequencies of the model, including through linearization or system identification-type simulations like white noise or free decay.
Best regards,
1 Like
Dear Dr. Jason,
1- I compared the outputs of OpenFAST (fora aft bending moment at mudline, blade tip deflection, blade root moment) for monopile IEA 22 MW both in U.S. East Coast and Gulf of Mexico under their metocean data published by NREL, there is no difference between the loads (less than 2 % ), so the governing in that case is the wave excitation frequency, which will results in more stiffer tower with first bending natural frequency at least 0.2 compared to 0.15 in U. S. East Coast? I mean in this case it is not enough to optimize the turbine to be only between 1P rated +5% and 3Prated-5%, the governing for the design is wave excitation frequency which is site specific condition? Is my understanding correct?
2- is there a relationship between frequency, stiffness, and mass. I know that natural frequency is the square root of stiffness divided by the mass. Is there a way for increasing the stiffness while decreasing the mass at the same time? or there is no relation can be studied between stiffness and mass?
3- in my study, I model soil structure interaction as linear springs calculated from equations (depends only on the soil profile and pile inertia), the length is not included in these equation. is there a relationship between the mudline stiffness and the turbine stiffness? are theses stiffnesses are in series? which contributes most to the stiffness of the overall turbine? the SSI or the turbine to play around and try to decrease in mass with out violating the frequency constrain?
Sorry if my questions are not organized, I try to brain storm to find the factor that influence the most in the optimization.
Regards,
Marwa
Thanks, Dr. Jason for your comments and detailed reply.
Regards,
Marwa
