Question about tower stuffiness and BModes

Dear Everyone

I am using “NRELOffshrBsline5MW_Tower_ITIBarge4” in FAST 7 and trying to modify the tower fore-aft stiffness. I found a variable “FAStTunr(1)” in the tower input properties file. it was described as “Tower fore-aft modal stiffness tuner, 1st mode” and set to 1.0 as default. I changed this value to 0.8 hoping to change the tower stiffness. However, some one told me that I can not change the tower stuffiness by doing this, instead, I need to change the mode shape of tower. Is he right?

If he is right, what do I actually change when I change the value of “FAStTunr(1)” to 0.8?

If I need to change tower mode shapes to actually change the tower stiffness, I know I need to use BModes and ModeShapePolyFitting.xls to calculate the 6th order polynomial coefficients. I tried to use BModes and get the output files shown below.

I think I need to extract the value of variable “f-a disp” of mode 1 and 2 in the output file and use it in the ModeShapePolyFitting.xls to get the coefficients. Is that right?

Best

XING WEI

Dear Xing Wei,

I would normally only use the tower stiffness tuners (FAStTunr(1) and (2) and SSStTunr(1) and (2)) for small adjustments e.g. to fine-tune the eigenfrequencies of individual tower modes to known solutions. By “small” I mean that I would not recommend setting the tuners outside the range, say, of 0.9 to 1.1. If larger adjustments are needed, this would suggest a problem with the specified tower distributed mass and/or stiffness data.

For sizeable changes to the tower stiffness e.g. from a redesign of the tower, you are correct that the mode shapes should be updated for consistency. Your understanding of BModes is correct–column “f-a disp” should be used for the fore-aft modes and column “s-s disp” should be used for the side-to-side modes.

I hope that helps.

Best regards,

Dear Jason

To redesign the tower, I think that I need to modify the value of TwFAStif TwSSStif TwGJStif TwEAStif TwFAIner TwSSIner of the tower at each point and then use BModes and ModeShapePolyFitting.xls to calculate coefficients. Is that right? Besides, I am wondering what else should I modify like “tower f-a mass moment of inertia about the tip-section y reference axis” as I want to change the stiffness. Because I do not know if there is any other influence on other variables as I change the stiffness of tower.

Best

XING WEI

Dear Xing Wei,

Yes, that is correct except that if you are only using FAST v7, and not the FAST-to-ADAMS preprocessor, the only columns of data used in the tower file are TMassDen, TwFAStif, and TwSSStif (TwGJStif etc. are only used by the FAST-to-ADAMS preprocessor).

Best regards,

Dear Jason

Your reply helped a lot! I am editing main input files of BModes to modify the mode shape of tower and got some problems. I am trying to modify the variables shown below.

I used NRELOffshrBsline5MW_Tower_ITIBarge4 in FAST 7 and I suppose that the tower top mass equals the mass of hub, nacelle and blades which can be found in the pdf “Definition of a 5-MW Reference Wind Turbine”. I can not find the mass of rotor in pdf so I am wondering if it has been included in the mass of nacelle?

For the tower top mass c.m. offset, I used the c.m. offset of nacelle. However, I am not sure if it is accurate enough because I am wondering if the mass of hub and blades have been considered together when they calculate the c.m. offset of nacelle?

Besides, I think the variable “izz_tip”, described as the torsion mass moment of inertia about the tip-section z reference axis is actually is the nacelle inertia about yaw axis which can be found in pdf. Is that right?

For the other variables shown in picture above, where can I find these values or do I need to calculate them? How to do the calculation if I need to?

Best

XING WEI

Dear Xing Wei,

The tower-top mass/inertia properties of the NREL 5-MW turbine are documented in the following forum topic: Using Aggregate Mass in ADAMS to Check NREL CS_Monopile.bmi.

Best regards,

Dear Jason

I thought you are talking about the values of moment of inertia in the reply shown below.

If so, I think I need to use the value of inertias of the RNA(rotor-nacelle assembly) specified relative to the CM of the RNA which is shown below.

Ixx = 4.370E7 kgm^2
Iyy = 2.353E7 kg
m^2
Izz = 2.542E7 kgm^2
Izx = 1.169E6 kg
m^2

and the value of IXY and IYZ shown in the upper part of picture?

By the way, from the primary input file of Fast 7. I found that the value of “NacCMxn”, “NacCMyn” and “NacCMzn”(described as the Downwind, Lateral, Vertical distance from the tower-top to the nacelle CM ) do not match with the locations you provided in the picture. Why? Is it because these three variables means differently with the locations you provided or other reason?

By the way, I think the values of “ixx_tip”, “iyy_tip”, “izz_tip”, “ixy_tip”, “ixz_tip”, “iyz_tip” do not depend on the type of tower or if it is offshore, if there is a platform. It only depend on the properties of hub, blades, and nacelle so these values keep the same for NREL 5.0 MW Baseline Wind Turbine.

Please correct me if I am wrong.

Best

XING WEI

Dear Xing,

Yes, your understanding is correct.

Ixy and Iyz are zero for all practical purposes.

FAST input paraemters NacCMxn, NacCMyn, and NacCMzn apply only to the nacelle, not including the rotor (hub and blades), which is why they differ from the CM reported for the RNA.

Best regards,

Dear Jason

Finally, I run BModes and use ModeShapePolyFitting.xls to calculate the mode shapes of NREL-5MW Baseline wind trubine. However, I found that the coefficients of mode shapes provided by “NRELOffsheBsline5MW_Tower_ITIBarge4” are different from what I calculated.

I used the values shown below in BModes main input file:

False Echo Echo input file contents to *.echo file if true.
2 beam_type 1: blade, 2: tower (-)
0. romg: rotor speed (rpm), automatically set to zero for tower modal analysis
1.0 romg_mult: rotor speed muliplicative factor (-)
87.6 radius: rotor tip radius measured along coned blade axis OR tower height (m)
0. hub_rad: hub radius measured along coned blade axis OR tower rigid-base height (m)
0. precone: built-in precone angle (deg), automatically set to zero for a tower
0. bl_thp: blade pitch setting (deg), automatically set to zero for a tower
1 hub_conn: hub-to-blade connection [1: cantilevered; other options not yet available]
10 modepr: number of modes to be printed (-)
f TabDelim (true: tab-delimited output tables; false: space-delimited tables)
f mid_node_tw (true: output twist at mid-node of elements; false: no mid-node outputs)

--------- Blade-tip or tower-top mass properties --------------------------------------------
350000 tip_mass blade-tip or tower-top mass (see users’ manual) (kg)
0.413 cm_loc tip-mass c.m. offset from the tower axis measured along the tower-tip x reference axis (m)
4.370E7 ixx_tip blade lag or tower s-s mass moment of inertia about the tip-section x reference axis (kg-m^2)
2.353E7 iyy_tip blade flap or tower f-a mass moment of inertia about the tip-section y reference axis (kg-m^2)
2.542E7 izz_tip torsion mass moment of inertia about the tip-section z reference axis (kg-m^2)
0.201 ixy_tip cross product of inertia about x and y reference axes(kg-m^2)
1.169E6 izx_tip cross product of inertia about z and x reference axes(kg-m^2)
2.88 iyz_tip cross product of inertia about y and z reference axes(kg-m^2)

--------- Distributed-property identifiers --------------------------------------------------------
1 id_mat: material_type [1: isotropic; non-isotropic composites option not yet available]
‘NRELOffshrBsline5MW_Tower_ITIBarge4_Properties.dat’ sec_props_file name of beam section properties file (-)

Property scaling factors…
1.0 sec_mass_mult: mass density multiplier (-)
1.0 flp_iner_mult: blade flap or tower f-a inertia multiplier (-)
1.0 lag_iner_mult: blade lag or tower s-s inertia multiplier (-)
1.0 flp_stff_mult: blade flap or tower f-a bending stiffness multiplier (-)
1.0 edge_stff_mult: blade lag or tower s-s bending stiffness multiplier (-)
1.0 tor_stff_mult: torsion stiffness multiplier (-)
1.0 axial_stff_mult: axial stiffness multiplier (-)
1.0 cg_offst_mult: cg offset multiplier (-)
1.0 sc_offst_mult: shear center multiplier (-)
1.0 tc_offst_mult: tension center multiplier (-)

--------- Finite element discretization --------------------------------------------------
12 nselt: no of blade or tower elements (-)
Distance of element boundary nodes from blade or flexible-tower root (normalized wrt blade or tower length), el_loc()
0. 0.08 0.16 0.24 0.32 0.40 0.48 0.56 0.64 0.72 0.80 0.90 1.0

--------- Properties of tension wires suporting the tower --------------------------------
0 n_attachments: no of wire-attachment locations on tower, maxm allowable is 2; 0: no tension-wire support (-)
3 3 n_wires: no of wires attached at each location (must be 3 or higher) (-)
6 9 node_attach: node numbers of attacments location (node number must be more than 1 and less than nselt+2) (-)
0.e0 0.e0 wire_stfness: wire sifnness in each set (see users’ manual) (N/m)
0. 0. th_wire: angle of tension wires wrt the tower axis at each attachment point (deg)

The red lines are the modification to original input file because I want to use the properties of NREL-5MW Baseline wind turbine and could you please help me to check if it is right. The values of “ixx_tip”, “iyy_tip”, “izz_tip”, “ixy_tip”, “ixz_tip”, “iyz_tip” are copied from my last post. Is it right to do so?

The section properties of ITI Barge 4 tower is attached and I believe it is correct. (I modified the file type from .dat to .txt so that it could be attached).

The outputs I got from BModes are shown below.

I am wondering why the f-a disp are all zero for mode 1? Because I can not get any result if I put zeros into ModeShapePolyFitting.xls to calculate mode shape coefficients.

By the way, I tried to modify “ixy_tip” and “iyz_tip” to some number similar to “ixx_tip” and then, I got some values for both s-s disp and f-a disp for all modes in output file. Why is that?

Best

XING WEI
NRELOffshrBsline5MW_Tower_ITIBarge4_Properties.txt (2.13 KB)

Dear Xing Wei,

Here are a few comments that have been discussed several times on this forum:

  1. Tower mode shapes are strongly effected by the tower-base boundary condition. That is, the mode shapes of a cantilevered tower will be different from the mode shapes of a tower mounted atop a floating platform. You’ll need to specify the platform mass/inertia, added mass, hydrostatic, and mooring stiffness in BModes if you wish to derive the tower mode shapes for floating wind system–see: BModes : Input parameters about tower support subsystem.
  2. The tower model of the ElastoDyn module of FAST v8 is simpler than that of BModes, which will effect how the distributed tower properties should be specified in BModes–see: Consistency between Bmodes and MBC - #2 by Jason.Jonkman.
  3. When the NREL 5-MW models were first developed, we used MSC.ADAMS to derive the tower mode shapes, not BModes, so even if the BModes model is assembled correctly, the results will differ somewhat from what NREL has provided–see: Generate new mode shapes for the 5MW ITIBarge4 wind turbine - #2 by Jason.Jonkman.

From your results, mode 1 is a side-to-side mode, uncoupled from fore-aft motion, and mode 2 is a fore-aft mode, uncoupled from side-to-side motion.

Best regards,

Dear Jason

Actually, the BModes input file I used does not have the lines to enter platform mass/inertia, added mass, hydrostatic, and mooring stiffness so I think the results I got is for the non floating wind turbine which does not have a platform. I suppose that I can not compare the coefficients of the mode shape I got with that of “NRELOffsheBsline5MW_Tower_ITIBarge4.dat” provided on your website as this one considers the platform. Am I right?

If so, I just want to deal with the non floating wind turbine and which file should I use to compare my result to see if I did the right calculation?

Best

XING WEI

Dear Xing Wei,

Yes, that’s correct.

The land-based (fixed) model of the NREL 5-MW turbine is provided as Test18 in the FAST CertTest in the FAST archive. The corresponding tower file is named NRELOffshrBsline5MW_Onshore_ElastoDyn_Tower.dat.

Best regards,

Dear Jason

I have successfully got the mode shape coefficients by BModes and ModeShapePolyFitting.xls. The coefficients of first f-a and s-s mode shape are not similar to the coefficients provided by “NRELOffshrBsline5MW_Onshore_ElastoDyn_Tower.dat” but the curves look like almost the same. But for the second f-a and s-s mode shape coefficients, their curves are different dramatically. I am wondering if I can directly use the mode shape coefficients obtained by BModes and ModeShapePolyFitting.xls?

By the way, Fro the frequency of 1st and 2nd f-a and s-s modes. they a little bit different from the values provided by “Definition of a 5-MW Reference Wind Turbine for Offshore System Development.pdf”. I am wondering if it because I derived these frequencies by using an onshore wind turbine tower file? The BModes input file I used does not include the information of “Platform-reference-point-referred hydrodynamic 6X6 matrix (hydro_M)”.

Besides, I am wondering if the TMD damper and mass are sensitive to the frequency as I changed the tower stiffness a lot of times to change the frequencies of 1st f-a and s-s modes from around 0.3 Hz to 0.15 and 0.17Hz. Then I changed the mode shape coefficients accordingly. However, I found that the TMD spring and damper could still control the displacements of tower in an effectively. The only difference is the final equilibrium point for different frequencies. I do not know why as I supposed that the spring and damper are sensitive to frequencies.

Best

XING WEI

Dear Xing Wei,

Yes, you can use BModes and ModeShapePolyFitting.xls to calculate tower mode shapes for FAST. When comparing your results to the existing tower mode shapes provided for the NREL 5-MW, I would expect some differences resulting form point (3) in my Feb 24, 2017 post above. I would expect the first tower mode shapes to be close (but not exact) and the second tower mode shapes to differ more as the compliance/flexibility in the drivetrain and rotor has more an affect on the second tower mode shapes. I wouldn’t expect much effect in the FAST simulation results, but I would be curious if you see any noticeable affect as a result of these changes to the tower mode shapes.

The natural frequencies in the NREL 5-MW specifications report you mentioned are reported for the land-based system, so, your frequencies should match closely. Again, the effect of compliance/flexibility in the drivetrain and rotor would have a stronger effect on the second tower natural frequencies than they do on the first i.e. you should compare to the natural frequencies derived from FAST rather than from BModes.

I cannot really comment as to how sensitive the optimal TMD properties are to changes in the tower design.

Best regards,

Dear Jason

So according to what you said in your last post, the reason why the second tower mode shape coefficients derived by BModes are different from what exists in FAST now is because MSC.ADAMS was used to obtain the coefficients in FAST which considered the compliance/flexibility of drive train and rotor. In BModes calculation, we did not consider the compliance/flexibility of drive train and rotor. Is my understanding right?

Besides, I think that if I change the tower stiffness at each point, the mode shape coefficients should be changed accordingly so natural frequencies change automatically because the change of mode shape coefficients. Is my understanding right? If it is right, I am wondering if there will be a big difference in simulation (like f-a displacements under same wind condition) when I change the tower stiffness, mode shape coefficients and frequencies in theory? Actually I want to figure out how does the natural frequency influence the simulation results (under constant wind speed or turbulence).

Best

XING WEI

Dear Xing,

Yes, your understanding is correct.

The natural frequencies will influence the response because the structure will respond to wind loading differently with changes to the natural frequencies. How much the response will change depends on what is changed, but you can certainly use FAST to answer that question.

But I’m not sure I understand your question, “I am wondering if there will be a big difference in simulation (like f-a displacements under same wind condition) when I change the tower stiffness, mode shape coefficients and frequencies in theory?”

Best regards,

Dear Jason

Thank you very much for your reply! What I want to say in last post is as followed. I am wondering if there will be some difference on tower top fore-after displacements before and after I change the mode shape coefficients and natural frequencies? As the natural frequencies and mode shape coefficients are obtained by BModes by changing tower stiffness, the question is actually about the influence of stiffness variations.

Thank you very much for your reply! Nowadays, I am trying to use FAST 6 to simulate the situation of onshore wind turbine. As you know that FAST 6 was by default to simulate the NREL 5-MW Reference Wind Turbine for Offshore and equipped with “NRELOffshrBsline5MW_Tower_ITIBarge4.dat” as tower file which means the wind turbine is offshore. However, I found that there is a variable “PtfmModel” to choose the platform type in the primary input file. 0 means that there is no platform, 1 means onshore, 2 means fixed bottom offshore and 3 means floating offshore . It was set to 0 by default.

I am wondering what is the difference between 0 and 1 because I think there will be no platform if the wind turbine is onshore.

Besides, I am wondering if I want to simulate an onshore wind turbine, I need to use the “NRELOffshrBsline5MW_Onshore_ElastoDyn_Tower.dat” provided by FAST 8 as tower file and set “PtfmModel” to 0 . Is my understanding right? If not, what should I do?

Moreover, I am wondering what it means if I set “PtfmModel” to 0 and use “NRELOffshrBsline5MW_Tower_ITIBarge4.dat”. I noticed that the mode shapes coefficients are different between “NRELOffshrBsline5MW_Tower_ITIBarge4.dat” and “NRELOffshrBsline5MW_Onshore_ElastoDyn_Tower.dat” so I am a little confused by which one should I use.

The reason why I do not use FAST 8 is because I need to use FAST-SC which is base on FAST 6.

Best

XING WEI

Dear XING WEI,

Yes, I would expect some influence of tower-top displacements as a result of changes to the mode shape coefficients and/or natural frequencies.

I’m not sure why you say that the onshore NREL 5-MW turbine would use the NRELOffshrBsline5MW_Tower_ITIBarge4.dat file. The onshore NREL 5-MW turbine should use the NRELOffshrBsline5MW_Tower_Onshore.dat file. While the distributed mass and stiffness are the same for the two towers, the mode shapes differ because the tower-base boundary conditions differ. The old FAST models of the NREL 5-MW turbine in the onshore and various offshore configurations are available here: wind.nrel.gov/public/jjonkman/NR … Bsline5MW/.

In FAST v6 and FAST v7, PtfmModel = 0 ensured that the tower base was cantilevered to the inertia frame. PtfmModel = 1 caused the tower base to be cantilevered to the platform, which could have up to 6 degrees of freedom enabled e.g. for modeling soil-structure interaction for onshore turbines. In this case, the “platform” would represent e.g. the concrete block used in the foundation.

FYI - The functionality of FAST-SC is included in the TMD submodule of the ServoDyn module of FAST v8.

Best regards,

Hi Jason,

I am running the analysis of offshore wind turbine based on monopile substructure in OpenFAST v3.1. I know that tower properties in elstadyn are provided in terms of mode shape coefficients with the help of ModeShapePolyFitting.

I wanted to know that for every change in stiffness values at mudline (Text file) provided in Subdyn. Do we need to short run (say for 2sec) the OpenFAST for extracting the stiffness and mass matrix at tower base from subdyn sum file to BModes and then using ModeShapePolyFitting to get the coefficients? or is it fine if I don’t change the mode coefficients in elastodyn tower for every change in stiffnees values at mudline.

Thanks,
Satish J

Dear @Satish.Jawalageri,

Ideally you’d recompute the tower mode shapes in ElastoDyn with each change to the stiffness of the substructure. But it is difficult for me to guess if this would make much difference to the results in your case. I would suggest recomputing the mode shapes with the lowest and highest expected stiffness to see if that makes a difference; then decide for the remaining stiffness values.

By the way: You can also run standalone SubDyn through the SubDyn driver to compute the SubDyn summary file uncoupled from OpenFAST. This would prevent you from having to run a short OpenFAST simulation only to extract the SubDyn summary file.

Best regards,