Foundation Stiffness and damping in FAST v8

Dear Jason,

yes I run the simulation and get acceptable results with the workaround that you have recommended. I just wanted to know if the error is fixed and I get the error because I’m doing sth wrong while simulating with the potmod only. Now I know that it’s not fixed yet.

Thank you for your answer.

Best regards
Masih

Dear All,

I have read the previous posts in this topic and some others about modelling soil-structure interaction in FAST. My understanding is that for FAST v8 the only option currently available appears to be the apparent fixity approach.

I have read earlier that one option would be to modify the FAST source code itself. I was wondering if someone on the forum knows of anyone who attempted to do this, and whether they succeeded or not? I know this question is well beyond the scope of the help one should expect from this forum, but how difficult would it be to modify the source code for someone experienced with coding in Python and Matlab but only vaguely familiar with FORTRAN?

Alternatively, if I would try and use FAST v7 instead, in what ways is it inferior compared to FAST v8 when used for both monopiles and jackets?

Any help is greatly appreciated.

Kind regards,
Laszlo Arany

Dear Laszlo,

Your understanding is correct.

That said, NREL introduced a user-specified platform loading module (ExtPtfm) in OpenFAST, which mimics functionality previously available in the old UserPtfmLd() routine of FAST v7. See the Transition to OpenFAST and OpenFAST Release Notes for more information: wind.nrel.gov/nwtc/docs/README_OpenFAST.pdf.

Furthermore, NREL has developed a version of SubDyn that allows for a linear CS foundation (6x6 stiffness matrix) to be applied at the seabed instead of forcing a rigid connection between the substructure and seabed. However, this version needs to be reviewed, tested, and documented before it can be officially released.

Modeling more advanced foundations e.g. DS will require further development that will take a decent amount of work.

Certainly you can use the FAST v7 to model various foundations (AF, CS, and DS with linear or nonlinear springs), but please note that FAST v7 is limited to modeling wind turbines atop monopiles; it is not possible (without extensive modifications of the source code) to model tripods or jackets in FAST v7.

Best regards,

Dear Jason,

Thank you for your reply.

I have had a look at the “FAST v8 and the transition to OpenFAST” document, and my understanding is that OpenFAST v1.0.0 has all the functionality of FAST v816.00a-bjj, with some bug fixes and improvements implemented. Is this correct?

Is there a planned release date (or a rough idea) when the SubDyn version supporting linear CS foundations will be released?

Many thanks and kind regards,
Laszlo

Dear Laszlo,

Yes, OpenFAST v1.0 is basically an upgrade from FAST v8.16. The only thing missing from FAST v8.16 that is not yet in OpenFAST v1.0 is the Simulink interface–see the OpenFAST issue: github.com/OpenFAST/openfast/issues/59.

I would expect the upgrade to SubDyn supporting linear CS foundations would be made public this summer, but I don’t yet have an exact release date.

Best regards,

Dear Mr Jonkman,

My Bachelor exam deals with modeling a turbine-soil interaction with FAST. I recieved a running version of a 10MW Onshore Wind Turbine which I transferred to an OffshoreWT already.
Now, after finishing the calculations with a fixed towerbase at the mudline, I want to model the turbine-soil interaction with varieties of soil layers and stiffness parameters (those of OC3 studies) to evaluate the influence of the soil parameters on the OWT’s eigenfrequencies by analysing them.

The running model is based on FAST v.8 and I’ve already read a lot about the difficulties modeling the soil. Transferring the whole model to FAST v.7 (which could model a distributed spring model as I read) seems a little costly at the first glance, isn’t it? (1)

Another opportunity I read was to use FAST v.8 and HydroDyn only with platform additional stiffness and damping to imitate a soil layer. But here we’d just model 1 soil layer and no water, right? (2)

Then I read that there is a SubDyn upgrade upcoming, which supports linear coupled springs foundations. When could that preanounced update be available? (3)

The last option seems to be the Apparent Fixity like described here: http://forums.nrel.gov/t/aerodyn-feedback/57/1 (dialogue with Athanasios Kontis)
So this method can be useful to get different values for the mode shapes for different soil parameters, did I understand that correctly? (4)

What would you recommend me to do? One of my tasks is also to model aero- and hydrodynamic loads on the turbine with embedded monopile foundation. This is also possible in FAST v.8 through the AF model, isn’t it? (5)

Thank You all for this valueable forum, I’m really greatful!

Best Regards,

Dear Lasse,

Here are my answers to your questions:

(1) I guess that depends on how familiar you are with FAST v7. If you have only used FAST v8, then this may take a bit a work. If you’ve already upgraded from FAST v7 to FAST v8 (e.g. with other FAST models), then this is probably not too hard.

(2) Correct. This is for the case without SubDyn.

(3) I’m not sure. I believe Rick Damiani has a branch of OpenFAST with the linear foundation stiffness and damping implemented, but I’m not sure when he intends to submit a pull request to get this functionality into the development branch of OpenFAST. I’ll ask Rick about this when I can.

(4) The apparent fixity (AF) model mimics the effect of the soil-pile interaction on the dynamic response of the bottom-fixed wind system above the seabed i.e. you should be able to get the correct natural frequencies, deflections of the support structure, etc. if you know how to define the AF pile (length geometry, materials) correctly.

(5) Correct. The AF model is typically fine if you can define the AF model correctly and simply want the influence of the foundation flexibility on the dynamic response of the bottom-fixed turbine above the seabed, unless you specifically need the loads within the pile below the seabed, in which case the AF model is not sufficient.

I hope that helps.

Best regards,

Hello there,
I have run the FAST v.8 analysing a fixed-based OC4 jacket for DLC 1.2 (power production) recommended by IEC 61400-3 with the following configurations:
Hs=1,Tp=6 in HydroDyn and associated URef= 8 in Turbsim all regarding the scatter diagram presented in “Upwind Design Basis” for K13 deep water site in dutch North sea. I have been looking for shear forces and Moments at mudline at nodes which have been fixed to the seabed thus, used dynamics nodal forces requests at the end of SubDyn input file(i.e. M1N1FMxe,M1N1MMye).After finishing first simulation by setting first and second random seed numbers required by Turbsim and HydroDyn I have checked the result forces and moments time series while their amplitudes are not exceeded 4 (N) which is too small ! I was wondering whether it is related to the Hs value or not ?! Do you consider it as a right analysis ? Or you believe there is something wrong with it ?
Best regards,
Arsalan

Dear Arsalan,

I would guess the problem is that you are using the dynamic nodal loads rather than the static nodal loads. That is, M1N1FMxe and M1N1MMye should be changed to M1N1FKxe and M1N1MKye etc. The static nodal loads represent the reaction load within the member, not accounting for internal structural damping. The reaction load within the member is not well characterized by the dynamic nodal loads.

This point can be understand as follows for a simple single degree of freedom mass-spring damper system, which has the equations of motion:

mxdotdot + cxdot + k*x = Fext

The reaction load is

Freact = kx + cxdot = Fext - m*xdotdot

The static nodal loads output from SubDyn are equivalent to kx and the dynamic nodal loads are equivalent to mxdotdot. When calculating the reaction loads, the former only neglects the internal structural damping (which is usually small), while the latter neglects the applied loads (which are usually large).

I hope that helps.

Best regards,

Dear Jason,
I am really thankful for your answer.That helped me greatly.
Best regards,

Dear Mr Jonkman,

I really appreciate your answers, they helped me a lot.

I decided to go back to version 7 to model the DS system as the final task.

The transfer of the 10MW turbine from v.8 to v.7 isnt that hard I expected, but nevertheless I have a question about it:
Do the structural blade properties for FASTv.7 need the parameter “AeroCent” instead of “PitchAxis” or is it able to use “PitchAxis” as well? (If not,) is AeroCent = 0.25 - PitchAxis correct?

How do my next steps will look like? I have already read that I have to compile FAST on my own in order to use the UserPtfmLd() routine like described here: wind.nrel.gov/public/bjonkman/De … ctions.pdf

In the Chapter “Creating FAST with the User-Defined Control Options for Interfacing with GH
Bladed-style DLLs” is explained how to create an ownFAST_dll.exe where the UserPtfmLd() routine is implemented.
After that the chapter “Using Microsoft Visual Studio” begins. Is this necessary or does the ownFAST_dll.exe will be sufficient?

For the Analysis I will use the same soil parameters like in OC3 studies, but another Turbine than the NREL 5MW. Do I need to make changes in the published UserPtfmLd() file?

Further, I have trouble to use MBC3 and Matlab to get the eigenfrequencies of the entire system out of the .lin files. It’s because I didn’t use Matlab before. May you have a source where the application of Matlab for this or a similar case is described a little more detailed than in MBC3-User’sGuide?

I know those questions don’t fit to this topic but it seemed to me more rational to do it that way. Let me know if its not okay.

Best regards,
Lasse Meyer

Dear Lasse,

Here are my answers to your questions:

According to the FAST v8 ReadMe file (wind.nrel.gov/nwtc/docs/README_FAST8.pdf), in the conversion from FAST v7 to FAST v8, PitchAxis = 0.5-AeroCent. Thus when converting the other way, AeroCent = 0.5-PitchAxis.

Bonnie’s compiling instructions provide guidance for both compiling using the command line and compiling using Visual Studio. Which are you using?

The UserPtfmLd() routine is for specifying your own 6x6 mass, stiffness, and damping matrices. The UserTwrLd() is for implementing springs distributed along the tower/pile. You should implement/modify UserTwrLd() for a distributed springs (DS) model.

No other documentation has been developed, but the application of MBC3 has been discussed many times in this forum. What trouble are you running into?

Best regards,

Dear Mr Jonkman,

Thank you for the immediate answer.

Do I get it right: (?)
I have to model a monopile. Then I will use Software (LPile often mentioned, I will use IGTH Pile) to get springs’ stiffnesses out of this monopile and the soil parameters (of OC3 studies). I have to change the values for “zi” and “springs” in the UserTwrLd().f90 file. Are these the only necessary changes?

Then I also need to add the new towerfile_monopileDS.dat (where the monopile is included through the structural properties of the tower) to the fast7inputfile.fst.

What exactly is meant with the platform here? Is it the “transition piece”? My task doesn’t tell me about modeling it and I should compare my calculated eigenfrequencies with those out of literature to get no or explainable differences. I could imagine the platform will have a pretty strong influence on the eigenanalysis.
Is it possible to assume a rigid connection between the tower and monopile?

  • Update at this place - I think now I understand: This “UserPtfmLd()” will be used for modeling the CS-Model?

The other issues are clear now or I will specify the questions in appropriate topics.

Thank you very much for your help.

Best regards,

Dear Lasse,

Here are my answers to your questions:

Correct. The sample UserTwrLd() routine available in UserTwrLd_DS.f90 provided by NREL also calls the routine to calculate the Morison-based hydrodynamic loads, so, the added mass, drag cofficient, and diameter of the monopile are also defined within this sample UserTwrLd() routine.

When FAST v7 refers to the “platform”, it is is referring to the rigid body with six degrees of freedom that the tower-base is cantilevered to. When the tower + monopile is modeled as the “tower” in FAST v7, the platform then is the bottom end of the monopile. So in FAST v7, the platform is not the transition piece.

Correct.

Best regards,

Dear Jason,

Could I know Is it possible to specify the spring stiffness at the base of pile to account for soil structure interaction in OpenFAST instead of apparent fixity method.

Thanks,
Satish J

Dear Satish,

As of OpenFAST v2.6 and newer, it is now possible in SubDyn to model a 6x6 stiffness matrix of an elastic foundation by specifying an SSIfile at each base reaction joint. See the updated SubDyn documentation on readthedocs, as well as the sample input file here:
openfast.readthedocs.io/en/main … index.html
openfast.readthedocs.io/en/main … ndixA.html
openfast.readthedocs.io/en/main … ndixC.html

Best regards,

Dear Jason,

Thanks for your reply.

I am getting “fixed bottom case detected” even after specifying SSI file in Subdyn. Could you please advice me on this. I have attached Subdyn and SSI file. I am using OpenFAST v3.0.

Thanks,
Satish J


NRELOffshrBsline5MW_OC3Monopile_SubDyn.zip (3.69 KB)

Dear Satish,

I haven’t looked at your input file(s), but this is expected behavior. As of OpenFAST v2.6.0, SubDyn determines based on SubDyn input file settings whether the substructure is fixed-bottom or floating and treats some terms differently for floating systems (such as the use of a floating reference frame). The soil-structure-interaction (SSI) file is used for fixed-bottom systems. See the SubDyn documentation now on OpenFAST readthedocs for more information: openfast.readthedocs.io/en/main … index.html.

Best regards,

Dear Jason,

Thanks for your reply.

Does that mean Analysis includes the springs at the MP bottom instead of rigid foundation. If it is, could I please know where I can find that it is using stiffness values ( I cant see anything either in .SD.sum or .SD.echo). I did had a look at the documentation.

Thanks,
Satish J

Dear Satish,

“Fixed bottom” does not mean a rigid foundation; you can have soil stiffness matrix enabled. But I took a brief look at your input files and I see that you’ve specified the SSIfile correctly, but I also see that you haven’t enabled any of the degrees of freedom at the reaction joint. You should enable the DOFs at the reaction joint (by setting RctTDXss = 0, etc.) for the stiffness matrix to take effect.

Best regards,