I would like to change the dimensions of a tower and analyze it in FASTv7.
Therefore, we will modify Tower.dat.
How do you calculate TwFAM1Sh(2-6) and TwFAM2Sh(2-6)? (Using BModes?).
I would also like to know how to calculate TwEAStif.
(Can it be calculated with FAST?)
Dear @Kazuya.Nishiyama,
You should compute the tower mode shape polynomial coefficients by fitting a polynomial to the tower mode shapes.
As has been discussed in other forum topics, in terms of tools provided by NREL, you could use BModes (or BModes_JJ discussed on this forum if the tower is installed on a flexible foundation or offshore substructure), or Frame3DD, which is called within WEIS to compute tower mode shapes. BModes can also be used to derive mode shapes for spinning blades as needed by ElastoDyn for the blade mode shapes. You could also use other structural analysis software that can compute mode shapes not provided by NREL.
TwEAStif is the local value of the tower cross-sectional axial stiffness, E*A, as defined in the FAST v6 User’s Guide. But this input (and all columns after TwSSStif do not need to be specified unless you are using the FAST-to-ADAMS preprocessor functionality.
Best regards,
Thank you for your response.
I downloaded and checked BModes.
…1st question.
I tested the tower_sec_props.dat in test3 of BModes, which was the tower data for WP1.5MW.
It output 1-20Mode.
To find TwFAM1Sh, TwFAM2Sh, TwSSM1Sh, and TwSSM2Sh to be input to Tower.dat in FAST, is the following correct to be input to xy in ModeShapePolyFitting?
TwFAM1Sh(x=span_loc,y=f-a disp(Mode No.1))
TwFAM2Sh(x=span_loc,y=f-a disp(Mode No.2))
TwSSM1Sh(x=span_loc,y=s-s disp(Mode No.1))
TwSSM2Sh(x=span_loc,y=s-s disp(Mode No.2))
Second question
I have a question about Test3_tower.bmi in BModes.
I don’t understand the parameter for tower-top mass properties.
Please tell me.
tip_mass=RNA mass?
cm_loc=distance from hub center of gravity to tower center?
cm_axial=distance from tower top tip to nacelle center of gravity?
ixx_tip=TwFAStif(HtFract=1)?
iyy_tip=TwSSStif(HtFract=1)?
izz_tip=TwGJStif(HtFract=1)?
ixy_tip=?
izx_tip=?
iyz_tip=?
Dear @Kazuya.Nishiyama,
I agree with your definitions of TwFAM1Sh, etc.
The tower-top properties in BModes are the tower-top mass, center of mass, and second mass moment of inertias, all with respect to the tower-top location.
Best regards,
Dear Jason
Thank you for your answer.
I have four questions.
・Is it possible to calculate ixx_tip ～iyz_tip with FAST?
・I would like to know how to calculate flp_iner and edge_iner.
・I have a question about tower_sec_props.dat.
I do not know the values of tor_stff and axial_stff.
Is it possible to analyze without inputting them?
If values are needed, can they be calculated ?
・Is there a new manual for BModes?
(I would like a manual with the towers listed.)
I don’t have the .inp file and don’t know how to run BModes.
Dear @Kazuya.Nishiyama,
Regarding the tower-top mass, center of mass, and moments of inertia (BModes input ixx_tip, etc.), you can calculate these through a FAST / OpenFAST linearization analysis by only enabling the platform DOFs and only having mass in your ElastoDyn module pertaining to the rotor and nacelle, e.g., as discussed in the following forum topic: OpenFast 2nd order Linearization.
Regarding the setting of flp_iner, etc. in BModes when computing blade mode shapes for ElastoDyn, see my post dated March 18, 2015 in the following forum topic: NREL 5MW Blade Mode Shape. A similar process should be applied when using BMods to generate tower mode shapes for ElastoDyn, e.g., tor_stff should be set very high relative to the bending stiffness (100 to 1000 times more).
I’m not sure what you mean by a “new manual” for BModes. Which manual are you using?
Best regards,
Dear @Kazuya.Nishiyama,
Here are my responses:
Invalid numerical input for file “.\Suzu_GE1.5s_Tower.dat”.
The error occurred while trying to read line 8 of the tower distributed properties.
I’m not sure. As with any input file processing error, I would enable the Echo option to debug.
I see that ∙tor_stff is multiplied by 100-1000 and axis_stff is a very large value, but is a rough value sufficient?
Yes, a rough value is sufficient. The point is to set tor_stff and axial_stff high enough that the stiffness does not impact the results for the bending modes you care about, but not too high so as to cause numerical issues with the BModes solver. So, I would just scale the flp_stff and edge_stff by a factor of 100-1000 to get in the ballpark.
I am using the following manual.
However, this manual does not mention anything about towers.
There is an updated manual of BModes provided with the release of BModes v3.00 that describes the tower inputs, here: https://www.nrel.gov/docs/libraries/wind-docs/bmodes_v3-00-00.exe?sfvrsn=a1ac5f84_1&download=true.
Also, I would like to run this at the command prompt, but how do I type in the command?
(I downloaded the BModes, but there was no folder with .inp files.)
The command is as described in the manual excerpt you shared. There are example input files in the BModes v3.00 archive I linked to above.
Best regards,
Dear.Jason
Thank you for your response.
・I have found that setting ADAMSPrep=2 or 3 causes an error. I would like to know the cause of this.
・How do I set the tower weight to 0?
(Set HH=0?).
・The .lin file is output, but where is the mass stiffness matrix output?
Dear @Kazuya.Nishiyama,
Here are my responses regarding FAST v7:
I have found that setting ADAMSPrep=2 or 3 causes an error. I would like to know the cause of this.
You shouldn’t set ADAMSPrep = 2 or 3 unless you using FAST to generate an ADAMS model. If are you generating an ADAMS model, FAST requires additional inputs, such as additional columns of distributed blade and tower data and the ADAMS-specific input file.
How do I set the tower weight to 0?
(Set HH=0?).
You can’t set the tower weight to exactly zero (FAST will complain about unphysical data), but you can get around FAST’s physical checks by setting the tower weight to near zero by prescribing the tower flexible length to near zero (by setting TwrBsHt very close to TowerHt) and you can set the tower mass per unit length (TMassDen) to near zero.
The .lin file is output, but where is the mass stiffness matrix output?
Select MdlOrder = 2 in the FAST Linear input file to generate the mass and stiffness matrices.
Best regards,
Dear Jason
The mass and stiffness matrices could be calculated.
・Perhaps TwrBsHt is not present in FAST-v7.
Therefore, I am thinking of setting TowerHt and TMassDen to almost 0, and HH will be calculated accordingly.
Is it possible to do it this way?
・As stated in other topics, am I correct in understanding that Ixx - Iyz (BModes) are calculated from the calculated matrix using the parallel axis theorem?
・Checking the configuration file for the WP1.5MW model, TowerHt is larger than HH.
In my opinion, WP1.5MW will have a larger HH.
I also substituted HH into the formula and calculated it to be 85.84m.
Why does this value not match the HH for the WP1.5MW model?
HH=84.2876,TowerHt= 87.39 ,OverHang=-3.3,Twr2Shft=1.61,ShiftTilt=-5
HH=TowerHt+Twr2Shft+OverHang*SIN(ShiftTilt)]=89.29
Dear @Kazuya.Nishiyama,
Here are my responses regarding FAST v7:
Perhaps TwrBsHt is not present in FAST-v7.
Therefore, I am thinking of setting TowerHt and TMassDen to almost 0, and HH will be calculated accordingly.
Is it possible to do it this way?
I forgot that in FAST v7, the tower-base height is named TwrRBHt (not TwrBsHt, as it is in ElastoDyn within OpenFAST). Lowering the hub-height could cause other issues, depending on how you use your model.
A stated in other topics, am I correct in understanding that Ixx - Iyz (BModes) are calculated from the calculated matrix using the parallel axis theorem?
Correct.
Checking the configuration file for the WP1.5MW model, TowerHt is larger than HH.
In my opinion, WP1.5MW will have a larger HH.
I also substituted HH into the formula and calculated it to be 85.84m.
Why does this value not match the HH for the WP1.5MW model?
HH=84.2876,TowerHt= 87.39 ,OverHang=-3.3,Twr2Shft=1.61,ShiftTilt=-5
HH=TowerHt+Twr2Shft+OverHang*SIN(ShiftTilt)]=89.29
I agree with all of your math, except that TowerHt = 82.39 m (not 87.39 m), thus HH is correctly calculated.
Best regards,
Dear Jason.
Thanks for the answer.
Are you saying that it is not a good idea to make TowerHt and TwrRBHt approximately equal?
Dear @Kazuya.Nishiyama,
I’m recommending that you should make TwrRBHt approximately equal to TowerHt to reduce the tower mass to zero.
Best regards,
Dear Jason.
Thank you for your answer.
I referred to this forum.
(Tower fore-aft modes shapes - #9 by Jason.Jonkman)
I did a linear analysis and the RNA mass matrix is shown below.
The weight and moment of inertia are larger than the values given in the forum and I am trying to figure out why.
Since it was based on the WP1.5MW model and I did not make any major changes (just a few dimensions, no weight changes), I expected to get almost the same results.
However, the results were very different.
Is there something wrong with my settings?
*When I check the .fsm file, the RNA weight is 83670kg even though I did not change anything in the WP1.5MW model and did the analysis.
However, even if the weight is correct, I believe the moment of inertia is too large.
[mass matrix].
83670.00 0.00 0.00 0.00 5401000.00 0.00
0.00 83670.00 0.00 -5401000.00 0.00 -146900.00
0.00 0.00 83670.00 0.00 146900.00 0.00
0.00 -5401000.00 0.00 351500000.00 0.01 9393000.00
5401000.00 0.00 146900.00 0.01 350500000.00 -0.33
0.00 -146900.00 0.00 9393000.00 -0.33 1925000.00
Changes
・fst file
OverHang (change of several tens of cm)
NacCMx,y,z,n (change of several tens of cm)
Twr2Shft(change of tens of cm)
TowerHT(TwrRBHt+0.01)
TwrRBHt(20m shorter than WP1.5MW model)
※fsmを確認するとRotor Inertia=2953255.5kg-m^2
・AD File
HH(TowerHt+Twr2Shft+OverHang*SIN(ShftTilt))
・Tower file
TMassDen=0.01
TwFA,TwSSStiff=0.01
Dear @Kazuya.Nishiyama,
I’m not sure why but your tower-top mass properties are different than those of the WindPACT 1.5-MW baseline turbine model provided in the FAST v7 archive. E.g., you have NacMass = 52839 kg whereas Test13.fst has 51170 kg.
Best regards,
Thank you for your answer.
The manual and the fst file of the wind turbine model in the link below states Nacelle mass=52839kg.
Is that correct?
WindPACT Reference Wind Turbines
GitHub - IEAWindSystems/WindPACT-RWT: WindPACT Reference Turbines
Dear @Kazuya.Nishiyama,
I’m not sure why, but there are small differences between these references and the FAST model available in the FAST v7 r-test. This model is now over 20 years old, so, I’m not sure why the models differ.
Best regards,
Dear Jason.
Thank you for your response.
Linearization analysis was performed using the test14 file.
The results of the analysis are shown below.
・The value differs from the inertia described in this topic by about two orders of magnitude.Why is this?
・In test14 I have no damping on the blades and tower, is it better to have no damping in the linearized analysis?
・Is it possible to calculate natural frequencies with BModes?
(M:6×6)
7.81E+04 0.00E+00 0 0 6553000 -0.01
0.00E+00 7.81E+04 0 -6553000 0 -95800.00
0.0 0.0 78060.000 0.0 95800.0 0.00
0.0 -6553000.0 0.008 553100000.0 0.0 7941000.00
6553000.0 0.0 95800.000 0.0 551900000.0 -0.76
0.0 -95800.0 0.000 7941000.0 -0.8 1815000.00
m=tip_mass=78060.000
x_g=cm_loc=-1.227
y_g=0.000
z_g=83.948
z_g-TowerHt=cm_axial=1.558
Ixx=553100000.00
Iyy=551900000.00
Izz=1815000.00
Ixy=0.01
Iyz=-0.76
Izx=7941000.00
Converted to top tower
Ixx=552910460.38
Iyy=551782429.95
Izz=1697428.39
Ixy=0.00
Iyz=0.00
Izx=63059480850720.10
・fst
・blade
・tower
・linear.dat
・fsm
Dear @Kazuya.Nishiyama,
I haven’t reviewed all of your files and calculations, but I see that you’ve miscalculated Ixx converted to the tower top. The 6x6 mass matrix you derived from FAST v7 is relative to the platform reference at the base of the tower. So, you should use z_g instead of cm_axial in conversion of Ixx. Likewise for other inertia terms.
Best regards,
Dear Jason.
Thank you for your response.
・Is the following calculation method correct?
Ixx(TowerTop)=Ixx(Ptfm)-m(z_g^2)
Iyy(TowerTop)=Iyy(Ptfm)-m(cm_loc^2+z_g^2)
Izz(Tower Top)=Izz(Ptfm)-m(cm_loc^2)
Izx(Tower Top)=Izx(Ptfm)+m(cm_loc)(z_g)
