Ansys model tower

Dear Jason,

I have to do an Ansys model of the Test07 tower because I want to import it as mnf in an Adams model generated by Fast v7.

Before importing the tower in Adams as mnf I want to create an Ansys model that gives me the same eigenfrequency given by Fast linearization for the tower.

I was thinking about creating an Ansys model ( beam element type) with variable cross section ( ctube) along the tower height that are the same number as in Fast model (the same ad tower nodes number that I put in .fst file). I also think that to find the same frequencies in a modal analysis in Ansys and Fast linearization I have to have “consistency” between the two software in two aspects:

  1. Tower top mass and Inertia (that I model as a “mass21” with given mass and inertia moments in Ansys that should be the same as in Fast).

  2. I need to have the same TwFAStif, TwSSStif and TMassDen along the tower sections (I mean that this parameter at the tower “sections” in Fast should be the same of my Ansys model for the cross sections that I use).

I do not have tower parameteres such as diamterer, radius, width or density of this tower so I was thinking to solve the problem 2) this way:

I want to find Re (tower external radius) and Ri (tower internal radius) to use in Ansys cross sections that give me the same TwStif (FA=SS) and the same TMassDen. So I have two unknowns that I find thanks to this equations for each tower section:

TwStif ( From Fast) =EI=E(pi/64)(D^4-d^4)
TMassDen (Fom Fast) =ro
A=ro*(pi/4)*(D^2-d^2)

So for each Ansys cross section I will use D and d calculated this way and put this in my Ansys beam189 model. This method uses the hypotesis that the density and Young modulus are constant along the tower section so I can write the integrals simply as roA and EI. And I suppose that the tower material il steel so that
ro=7500 [kg/m^3] and E=2.1*10^11 [N/m^2]. Do you think that this method is ok or there is an easier way to reach my goal?

Thank you very much!

Dear Alessandro,

I’m not an ANSYS user, so, I don’t know if there are alternative ways of defining a beam element in ANSYS.

But your approach for deriving the tower external and internal radii at each cross section as a solution of two equations and two unknowns sounds reasonable to me.

Best regards,

Dear Jason,

Thank you for your answer.
I’ve tried my method for sections parameters and putting the top mass at the tower top I’ve obtained good results for Test07:

Fast says: 1.6 Hz for the 2 first frequencies and 10.5 Hz for the 2 second frequency.

With my method instead I have 1.8 Hz and 11 Hz BUT only putting the top mass in Ansys. I would like to put on Ansys also the inertia moments but I have only many of them reading the fast .fst file like NacIner, GenIner and HubIner. I don’t have other Inertia moments ( for example Nac Iner about the FA or SS direction because Fast does not give them to me. I would like to say if FAST uses ALL the inertia moments ( that I can’t read in .fst file but It computes them inside) or It only uses these Inertia moments…because to make a comparison in modal analysis I should have “consistency” between moments as I said in my previous post.

thank you!

Dear Alessandro,

How are you obtaining these natural frequencies out of FAST e.g. through a FAST linearization analysis or by post-processing time series?

FAST does not directly compute the total tower-top inertia i.e. with the rotor, drivetrain, and nacelle treated as a single rigid body, but the effect of all the inertias and distributed masses are included in the dynamic response. Other than a laborious hand calculation, we’ve used two methods to derive the total tower-top mass and inertia by calculating 6x6 rigid-body mass matrix of the tower-top. You’ll find several examples on the forum where this has been done using either (1) the linearization of FAST v7 with only the six platform DOFs enabled (setting the tower and platform masses and inertias to zero to leave only the tower-top mass and inertias remaining) or (2) the FAST-to-ADAMS preprocessor of FAST v7 to generate an ADAMS model, from which the aggregate mass can be calculated within MSC.ADAMS. However, without modification of the source code, it is not possible to directly output the 6x6 mass matrix from FAST v8.

Best regards,

Dear Jason,

Thank you for your answer.
To obtain the natural frequencies I 've used a Fast v7 linearization with RotSpeed=0, no Aerodynamic forces calculated and only tower DOFs activated.

I’m not sure I’ve fully understood your answer but you let me think about generating an Adams model of test07 from Fast and than use the moments of inertia of the parts like generator, hub and all the others and to use these moments in my Ansys model.Is it ok in your opinion? So if I have the same section parameters, masses and moments of inertia I should see the same frequencies.is it right?

Dear Alessandro,

Ah, OK. In this case, FAST is essentially treating the total tower top (rotor, drivetrain, nacelle) as a rigid body in the linearization analysis. I would encourage you to use option (1) because you are already familiar with the FAST v7 linearization process and to avoid the complexity of using MSC.DAMS. That is, set the tower and platform masses and inertias to zero in the FAST input files and linearize the FAST v7 model with only the six platform DOFs enabled using MdlOrder = 2. The resulting 6x6 mass matrix will then represent the rigid-body of the total tower-top (rotor, drivetrain, nacelle) about the platform reference point, including the total mass (upper-left quadrant), center of mass (upper-right and lower-left quadrant) and inertias (lower-right quadrant).

I hope that helps.

Best regards,

Hi Jason,

I didn’t understand what you meant by “the linearization of FAST v7 with only the six platform DOFs enabled (setting the tower and platform masses and inertias to zero to leave only the tower-top mass and inertias remaining)”. I didn’t understand which parameters should be zeroed. Is it PtfmSgDoFs, PtfmSwDOf, etc.? I am not getting the same results as you have discussed in another forum, but rather a very higher range of mass and COM. What should be the wind parameter (wind speed)? and water-related parameters (wave height, water depth)?
Also I am trying to determine the mass moment of inertia of a tower, do I need to input the control input (CntrlInpt) and wind input disturbance settings (Disturbnc)? I did 0 in my linearization, is it okay?

Thank you for helping me out so far, I really appreciate that.

Regards,
Jharna

Dear Jharna,

If all your interested is the mass, center of mass, and inertia, then only the mass matrix resulting from the FAST v7 linearization calculation is important. The settings of wind, waves, control inputs, wind-input disturbances, etc. shouldn’t matter. Simply disable all DOFs except the platform DOFs (PtfmSgDOF = PtfmSwDOF = PtfmHvDOF = PtfmRDOF = PtfmPDOF = PtfmYDOF = True), set all of the masses you don’t want to zero (e.g. PtfmMass = PtfmRIner = PtfmPIner = PtfmYIner = 0), and set the masses you do care about nonzero. Linearize about the initial conditions (CalcStdy = False) and look only at the resulting 6x6 mass matrix, which will contain the mass, center of mass, and inertias of the remaining mass about the platform reference point.

Best regards,

Thank you Jason.
Now I have run the BModes and have the modal frequencies. I also used ModeShapePolyFitting.xls file to get 6th Order polynomial.
I was wondering if those coefficients should be the input in tower mode shape section of the elastodyn input file? This is because the SS2 and FA2 coefficients are much higher than that are available in elastodyn input file for 5MW Offshore monopile tower.
I am also modifying the TMassDens, TwFastiff, and TwSSStif column of the elastodyn file, I am not able to find those parameters during the linearization process. Could you please redirect me to the page if this had been discussed before?

Thanks,
Jharna

Dear Jharna,

I’m not sure what you are trying to do, so, it is difficult for me to comment. However, if you change the tower design or tower boundary conditions (mass/inertia of the rotor and/or nacelle, or substructure geometry/mass/stiffness), then you should update the mode shapes in the ElastoDyn tower file.

The tower distributed mass and stiffness (TMassDens, TwFAStfif, TwSSStif) are not found from a FAST linearization analysis. Instead, these should be based on the tower design (e.g. diameter, thickness, density, and elastic modulus).

Best regards,

Dear Jason,

I am testing on the NRELOffshrBsline5MW_Onshore (test18 of FAST v8.16). I am trying to adjust the tower sections, so I check the ElastoDyn file first with BModes. The first order mode shapes fit well, but The second mode shape I got from BModes was quite different from that defined in the ElastoDyn file. The tower section property file, .bmi file and mode shapes comparison are attached. I am wondering if there is something wrong in my data and if the BModes only is adequate for the mode shapes defined in the ElastoDyn file.

Best regards,

The mode shapes comparison:
First order mode shape
Bmodes ElastoDyn
x y1 y2
0 0 0
0.1 0.010288647 0.012368449
0.2 0.041326852 0.045509677
0.3 0.093376114 0.096855792
0.4 0.166589707 0.166705715
0.5 0.260852642 0.256498438
0.6 0.375598102 0.367470547
0.7 0.509600367 0.499698016
0.8 0.660744212 0.651522253
0.9 0.825770791 0.819360417
1 1 0.9979

Second order mode shape
Bmodes ElastoDyn
x y1 y2
0 0 0
0.1 -0.039552739 -1.047325878
0.2 -0.147348911 -3.646088512
0.3 -0.301508731 -7.125000822
0.4 -0.472989234 -10.84796621
0.5 -0.625094943 -14.10034375
0.6 -0.71197269 -16.07301677
0.7 -0.676090584 -15.94426474
0.8 -0.444701134 -13.05943859
0.9 0.074711481 -7.208439318
1 1 0.999
tower_sec_props.txt (2.12 KB)
Onshore.txt (4.01 KB)

Dear Yuqi,

I would not expect a perfect match to the mode shapes in the “NRELOffshrBsline5MW_Onshore_ElastoDyn_Tower.dat” file because I derived those mode shapes using ADAMS (with a flexible rotor) in place of BModes, as discussed in the following forum topics:
*http://forums.nrel.gov/t/nrel-5mw-onshore-using-bmodes-modeshapepolyfitting-xls/1106/8
*http://forums.nrel.gov/t/bmodes-input-parameters-about-tower-support-subsystem/604/1
*http://forums.nrel.gov/t/tower-eigenfrequencies-of-nrel-5mw-turbine/517/1

I see that you are using BModes v1.03 rather than the alpha version (BModesJJ). You may want to see if upgrading to the alpha version provides better results: nwtc.nrel.gov/BModes.

Best regards,

Dear Jason,

Many thanks for your detailed answer. I will try the alpha version. I hope it would provide better results and the difference of the second order mode shape wouldn’t lead to a big difference on the load.

Best regards,