Turbine-soil interaction. Influence on the mode shapes.

[Marwa.Mohamed]

Dear Dr. Jason,

I have one more question please.
If I want to study the effect of water depth, like 16m -25m -41 m for monopile IEA 22 MW, in this case I will change the water depth in HydroDyn

should I change the substructure part in SubDyn to be the same as water depth 16m, 25m, 41 m? or I can keep it 34 m as the manual for all of them? but there will be problem for 41 m, wright?

When I make the substructure length is 30m for cases (10m , 20m, 30m), the difference in response between the three water depths was around 18%

What is the correct approach for considering water depth variation?

Regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

If you want to change the water depth, in addition to WtrDpth, you should change the JointZss (or add joints) in SubDyn such that the lowest joint is at the revised seabed. Likewise, you should change the Jointzi or (add joints) in HydroDyn such that the lowest joint is at the revised seabed. This may require a redesign of the monopile because its change of length.

Best regards,

[Marwa.Mohamed]

Hello again,
I have another question please and it would be appreciated if you could help me with any comments.
For turbine IEA 15 MW on monopile, I modeled the soil pile system as coupled springs through SSI.txt file. I compared the effect of foundation flexibility considering different DLCs. There is no optimization here, I want to study the effect of SSI
as shown in the figure, some DLCs, the effect of SSI is minimum like DLCs (1.1,1.3,1.5,1.6,2.1,2.3), for DLC 5.1 around 30%, DLCs 6.1 and DLC7.1 is maximum effect around 90% and 80%, while DLC6.2 and 6.3 the trend is the opposite, it should be noted that DLC 6.2 and DLC 6.3 the results are taken from yaw error (20-60) which has very high post-stall angles-of-attack outside of the verified/validated linear regime and are not to be trusted as conventional numerical models that rely on 2D airfoil polars and 1D beams struggle to reproduce reality, which is characterized by unsteady aerodynamics and 3D effects because a the AeroDyn’s induction or dynamic stall models were not enabled in these DLCs. what was surprising that when I tested DLC 6.2 under yaw error 160 degree and wind wave are codirectional, the trend was the same as 6.1 which has yaw error 8 degrees and DLC 7.1, My questions are

1- Parked idling DLCs (6.1, 7.1, “6.2 with yaw error outside range ±20 to 60”) the SSI effect is higher than operation DLCs as the aerodynamic damping is least in these DLcs?
2- why DLC 5.1 has the maximum effect in operating DLCs? is it due to emergency shutdown so after shut down the aerodynamic damping is minimum? but why this does not happened with DLC 2.1 and 2.3 as they stopped also at specific time?
3- why DLCs 6.2 and 6.3 with these DLCs the trend is the opossite? is it like putting the structure on spring , so it absorb the energy?
4- with foundation flexibility, should the load increased due to (p- delta effect) secondary moment effect? or should it decreased because it is like put the structure on springs and it absord the energy so the moment decrease?

Regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

I doubt I can comment on your specific results without knowing a lot more about your set-up and results.

I would just say I would expect the biggest impact of the SSI on the global response is from the change to the support structure natural frequencies (moving closer or further away from excitation frequencies) and damping. So, when you change from fixed foundation to a foundation with SSI, I would suggest comparing the impact of SSI on the support structural natural frequencies and damping.

Best regards,

[Marwa.Mohamed]

Thanks Dr. Jason for your quick reply.
kindly find the attached picture of PSD for the three base conditions

as shown in this figure that as the soil type became more softer, the natural frequency decreased ( fixed base =0.195 Hz, stiff soil = 0.183 Hz, soft soil = 0.134 Hz).
but I don’t know why this effect changes from DLC to another? I mean if we design for example the turbine for only (DLC 101 and DLC1.3 (which is the case of IEA 22 MW decomentation) the effect of foundation flexibility is minimum, but when the same turbine experience another DLC like for example 6.1, it will have almost douple the load that it designed for, that’s why I want to study the influence on the DLCs.

Regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

Well, the excitation frequencies (from wind and waves) change with DLC, so, that is likely part of the answer. But to fully understand would likely require a deep dive into other response metrics associated with each DLC.

In addition to the change to the first-bending mode, I would also expend the second-bending mode to change, although this is a bit difficult to interpret from the PSD you shared.

Best regards,

[Marwa.Mohamed]

@Jason.Jonkman

he excitation frequencies (from wind and waves) change with DLC,

Thanks Dr. Jason for your quick reply.
1- you said that “the excitation frequencies (from wind and waves) change with DLC”, I think that’s why the effect of foundation flexibility is almost the same for DLC 1.1 and 2.1 which both have (NTM) and (NSS), but for DLC 2.3 it has (EOG), DLC 1.3 it has (ETM), and for DLC 1.6 it has (ESS). for DLC 5.1 and 6.1 and 6.3 maybe due to the aerodynamic damping is minimum?

2- what should I do to make the second bending mode more clear?

Regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

1. Could be.
2. I’m not sure how you are computing your PSD, but you can make PSDs more clear by binning within frequency bins and/or averaging across multiple time-series intervals.

Best regards,

[Marwa.Mohamed]

Dear Dr. Jason, thanks for your quick reply. I had another basic question please.
I read the paper of “Influence of Foundation Damping on Offshore Wind Turbine Monopile Design Loads” Carswell et al, 2021. they also investigated the effect of foundation flexibility in addition to damping considering different DLcs. as shown from the table, DLC 5.1 was the most DLC that has significant reduction in load after adding foundation damping

in addition, for DLC 6.1, they mentioned “The resulting small error in eigenfrequency using the criteria cited above was considered negligible for our analysis (it should be noted that the natural frequency from the representative mudline stiffness matrices was on average 0.23 Hz and varied by less than 0.01 Hz). The relative change in natural frequency can be seen in one example time history of tower top displacement (stochastic load case DLC 6.1a, refer to Section 4 and Figure 2A), where there is a significant phase and amplitude difference between the fixed base and first iteration flexible foundation case; with a second iteration, the difference in frequency and amplitude is substantially minimized. Methodology for the analysis herein relied on differences in foundation response: the difference in cyclic mudline moment amplitude from the fixed base to flexible base case was 39% but differed by only 2% after the second iteration (Figure 2B).”

my questions are.

1. How the natural frequency of the turbine changed from DLC to another (like what they have in DLC 6.1)? I thought that the natural frequency and mode shapes are depends only on the stiffness and the mass of the turbine and do not changes from DLC to another. Can you please explain this to me?
2. I have a similar response to that paper, where the foundation flexibility effects is maximum in DLCs 5.1 and 6.1. Is there any explanation?
   3)Should I exclude the results from DLC 6.2 and 6.3 with yaw errors between 20-60 from the results as they have the opposite trend?

Regards,
Marwa

[Marwa.Mohamed]

Dear Dr. Jason,

I want to add something more in paper “Effect of monopile foundation modeling on the structural response of a 5-MW offshore wind turbine tower” (jung el al. 2015)
he mentioned that"

he studied the turbine under DLC 1.1 rated wind speed. and in my results for DLC 1.1 the stiff soil difference is 0.15% and for stiff soil 4.6% similar to him (I studied turbines NREL 5 MW and IEA 15 MW and they both have the same trend).
I mean that when we studied the effect of foundation flexibility and soil effect in operation condition cases, the effect is not significant , but when we consider other DLCs such as DLC 5.1 and 6.1, the effect is more apparent.

1- Is my conclusion is correct please? and can you please give me more explanation.
2- for the second mode shape, I get the PSD from white noise in OpenFAST

Regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

I’m not familiar with the papers you are citing, and as I said, you’d have to dig into the details to fully understand the results you are showing. I can’t comment more on that.

Regarding (1), the natural frequency will depend on the mass and stiffness and boundary conditions. Depending on the nonlinearities of the model, these may change with magnitude of applied loads, which can vary between DLCs.

Regarding (3), I would not expect DLC 6.3 to have yaw errors larger than 20 degrees. This is certainly possible in DLC 6.2 and in the yaw error range of 20-60 degrees, I expect that you are running into the known blade edgewise / tower side-side instability that happens under parked/idling conditions under high winds and sizeable yaw error, which has been discussed in other topics on our forum:

Designing for yaw errors using FAST

[Yaw errors.png]Hello, I am trying to design monopiles for the NREL baseline 5MW turbine using site specific hurricane hazard according to IEC 61400-3 load case 6.2a, which requires the consideration of all yaw errors from -180 to +180 degrees. I have read some literature mentioning that there is an “instability” that occurs for the NREL baseline turbine at around +35 degrees. This is described as an “aero-elastic interaction causing negative damping in a mode that couples rotor azimuth with…

Wind Veer Parameter

Hi, I am trying to recreate the following study for a turbine at SWiFT: “Assessment of Wind Parameter Sensitivity on Ultimate and Fatigue Wind Turbine Loads: Robertson, Sethuraman, Jonkman, Quick (2018)”. I will be using TurbSim → OpenFAST and examining different parameters than those in the above study, but I did want to examine veer (β in the paper). I grepped all the input files and found nothing. I also didn’t find anything in the TurbSim Manual. How can I change the veer for these simulat…

Natural frequencies and mode shapes

Hello, I am currently working on a thesis’ project which includes the computation of campbell diagram, mode shapes and modal dampings of the blades of a 5MW NREL reference wind turbine. I would like to do so using the full model of the machine on OpenFAST. I have already managed to compute the power curve and the Cp-TSR curves of the rotor but I don’t know how to compute the natural frequencies, mode shapes and modal dampings at different wind speeds. Can you please show me how to do it if it’…

Tower strike with latest versions of OpenFAST and IEA 15MW

Hi forum, I am have been working with OpenFAST and the IEA 15MW RWT on the UMaine semisubmersible platform since May 2021, and this week, when I tried to update my software and model to v3.1 of OpenFAST and v1.1 of the turbine, my initial test simulations ended almost immediately with tower strikes. I wanted to report this issue and hear any advice others may have. For OpenFAST v3.1, I downloaded the Windows 64-bit executable from the OpenFAST GitHub releases page, and for the IEA 15MW model, …

High angle of attack at large wind speed

Hi, everyone I’m performing 10-min simulation with NREL 5MW turbine based on 5MW_Land_DLL_WTurb.fst with NTM condition in the wind speed range of 3m/s-25m/s. When the simulation is performed in large wind speed (e.g. 25m/s), I got the following warnings: FAST_Solution:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutput:RotCalcOutput:BEMT_CalcOut put(node 5, blade 1):UA_CalcOutput:UA_BlendSteady:Temporarily turning off UA due to high angle of attack or low relative velocity. This warnin…

wind-waves misaligned

Unable to reproduce the tower side-side instability in DLC 6.2 · OpenFAST...

[FOWT]Platform pitch and roll motions largely different from Bladed results under parked condition and specified azimuth angle.

Best regards,

[Marwa.Mohamed]

Thanks Dr. Jason for your comments. I really appreciate your help.

Regards,
Marwa

[Marwa.Mohamed]

Dear Dr. Jason,
I have a question please, and I hope you could help me.
1- I have optimized the structure considering the tower and monopile thickness distribution as well as the embedded pile length, with updates to the mode shapes, cross-sectional properties, and the soil-structure interaction (SSI) stiffness matrix. Following this optimization, the first natural frequency of the turbine was reduced to 0.136 Hz. The difference in dynamic loading before and after these updates is shown in the attached figure:

2- To further understand this behavior and investigate the potential for resonance, I decomposed the environmental loads into wind-only and wave-only cases. My objective was to determine whether the increase in loading was primarily driven by wave excitation, given that the optimized natural frequency is close to typical wave frequencies. The analysis showed that the load increase is primarily driven by wind rather than wave excitation:

However, the response in the X-direction still raised concerns:

my question is
1- 1. Is the observed increase in loading mainly due to resonance effects, or is it more a consequence of the increased flexibility of the optimized structure, which leads to higher deformations under wind loading?
2. Is it problematic that the optimized turbine has a natural frequency of 0.136 Hz, which lies within the range of wave excitation frequencies? Although wave-induced loading is not dominant, is there still a risk?
3. What are the implications of resonance in such a case, particularly considering that I have imposed constraints on von Mises stress and pile head displacements, and the optimized design satisfies those safety criteria? I am aware of the case presented in the report “Sustainable Installation of XXL Monopiles – Lessons Learned from SIMOX/COAX Offshore Campaign (Tsouvalas et al.)”, where resonance occurred during installation due to the vibro-hammer operating at a frequency matching the monopile’s natural frequency. This led to local stress amplification, material degradation, and eventual cracking and failure.

I would be very grateful for your insight into whether the observed behavior in my model indicates a structural risk or is an acceptable outcome given the design constraints applied.

[Jason.Jonkman]

Dear @Marwa.Mohamed,

It is hard to understand based only on the information you provided,but I would first figure out why the response of your model changes about 235-s into the simulation. Is this tied to some control action that kicks in at this time?

Regarding the implications of resonance, if an excitation force can excite a given mode and that mode is not heavily damped, it is generally best to avoid resonance other than for short periods of time.

Best regards,

[Marwa.Mohamed]

thanks Dr. Jason for your reply.
Yes, this case corresponds to an emergency shutdown event initiated at 230 seconds, which explains the observed change in structural response around that time. Given this, could the observed increase in loading be attributed to enhanced wave excitation and resonance effects, or is it primarily due to increased wind loading resulting from the structure becoming more flexible following the optimization? How can I know?

another question please, what is meaning of the mode is not heavily damped? does the damping changed from mode to another? how can I know this in the model based on the figures attached?

Regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

To answer your first question, I would suggest rerunning your wind-alone and wave-alone tests, but with the original design.

Each full-system mode has a natural frequency and a damping, both of which can depend on the operation of the turbine. OpenFAST can be used to calculate the natural frequencies and damping ratios of the full-system modes, typically expressed in a Campbell diagram as a function of rotor speed or wind speed. The Automated Campbell Diagram Code (ACDC) has been set up to automate the process of Campbell diagram generation using OpenFAST: GitHub - OpenFAST/acdc: ACDC: Automated Campbell Diagram Code.

Best regards,

[Marwa.Mohamed]

Thank you for your prompt response.

When I attempted to run the wave-only simulation on the original turbine model, I encountered an error when setting CompInflow = 0, as shown in the attached screenshot. Interestingly, this issue did not occur with the optimized turbine model.

However, when I changed CompAero = 0, the wave-only simulation ran successfully.

For your reference, I have included the results for the wind-only, wave-only, and combined load cases for the original turbine. Please note that for both the original and optimized models, soil–structure interaction has been accounted for using the SSI.txt file in the SubDyn module.

Best regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

With CompInflow = 0, you should disable various aerodynamic models in AeroDyn that are only valid with nonzero inflow; setting CompAero = 0 along with CompInflow = 0 is another solution, as you noticed.

Your results with the original turbine model imply to me that it is the aerodynamic excitation, not wave excitation, that are increasing the response for the optimized turbine.

Best regards,

[Marwa.Mohamed]

Dear Dr. Jason, thanks for your response.
So, Is the natural frequency of 0.136 Hz obtained for the optimized turbine considered problematic in terms of dynamic response or potential resonance effects? Or not?
Regards,
Marwa

[Jason.Jonkman]

Dear @Marwa.Mohamed,

Normally a support structure is designed as soft-stiff, with a first bending natural frequency between 1P and 3P across the operational speed range of the turbine. In your case, the low natural frequency will cross 1P in this speed range, which can lead to 1P resonance from rotor rotation.

Best regards,