Distributed Spring Model for Monopile.

Hello,

I am trying a monopile support structure in a 30 meter water depth and 45 meter soil-supported. I have following questions:

  1. In BModesJJ, which parameter is corresponding 45 meters soil-supported tower part.
  2. In BModesJJ, n_secs_k_distr: number of points at which distributed stiffness per unit length is specified (-)
    • Does this mean, the linearized spring can only be distributed along tower by unit meter?
  3. In BModesJJ, Distributed elastic stiffness per unit length along a flexible portion of the tower length:
    • Is this parameter the linearized spring stiffness which is the slope of the p-y curve when y is equal to 0? If not, what is the physical meaning of this parameter?
  4. Just for double check, in BModesJJ, is draft 30?

Last, did somebody get the 3 additional files mentioned in this Memorandum - Derivation and Description of the Soil-Pile-Interaction Models

Additional documents
OC3-Soil-Pile_InteractionModels.xls : Excel-Spreadsheet containing model data for the introduced foundation models and some validation data in terms of statics and dynamics.

OC3-Soil-Pile-Interaction Models_ReadMe.pdf :
Short note on the soil properties, pile properties as well as the soil-structure interaction models.

OC3-LPILE_Results.txt :
LPILE output file, containing data about the pile, soil, p-y-curves for some depths and some more calculated values.

If yes, can send me a copies of them. I could not find them from internet.

Really appreciate!!

Dear Lingling,

I have not used BModes myself to derive mode shapes for a support structure with a flexible foundation, but I’ll try to answer your questions:

  1. Input “draft” is the downward distance from the still water level to the base of the flexible support structure. So in your case, you’d set draft = 45.
  2. No, this means that the spring forces are calculated as forces per unit length along the tower. That is, the spring has the units of (N/m)/m. The first “m” comes from deflection “y” in a p-y model and the second “m” comes from the length of the tower element. So, the total spring force, dp, on an element of length dz is: dp = kydz, where “k” is the spring stiffness per unit length along the tower.
  3. Yes. See the answer to (2) above.
  4. No. See the answer to (1) above. The 30-m water depth implies that the spring acts over 45 - 30 = 15 meters. So, the distributed springs should be applied from 0 to 15 m above the base of the flesible support structure.

Please find the requested files attached.

Best regards,
OC3-LPILE_Results.txt (95.4 KB)
OC3-Soil-Pile-Interaction Models_ReadMe.pdf (39.7 KB)
OC3-Soil-Pile_Interaction_Model.xls (41 KB)

Dear Jason,

Really appreciate your files.

But, I still feel confused about the input “draft” - the downward distance from the still water level to the base of the flexible support structure.

In my case, water depth is 30m (the distance from mudline to MSL), the pile penetration depth is 45m (the pile length below mudline). If “draft” is referring to the distance from the still water level to the base of the flexible support structure then should this “draft” be 30 + 45 = 75m since this pile is flexible all the way down to the end.

Also, if I specify 45 for n_secs_k_distr: number of points at which distributed stiffness per unit length is specified (-) in BModesJJ, does this already mean the pile with distributed spring is 45 meters long?

Dear Lingling,

I’m sorry I misread your original forum post. I thought that you were referring to only 15 m of pile penetration. If the pile penetration is 45 m at a site with a 30-m water depth, than I agree that BModes input “draft” should be set to 75.

BModes input “n_secs_k_distr” is the number of sections where the distributed foundation springs are specified. That is, “n_secs_k_distr” is the size of the “z_distr_k” and “distr_k” arrays. If you wanted a constant or linear distribution of springs from the mudline to the bottom of the pile, you could set “n_secs_k_distr” to 2 and z_distr_k(1) to 0 and z_distr_k(2) to 45. If you want a nonuniform distribution of springs, you could add more sections.

I hope that helps.

Best regards,

Dear Jason,

Really appreciate.

Is the following the right way to do it? And also I am curious how did you deal with the mode shape input for FAST in OC3? Did you use ADAMS to generate the mode shapes?

Distributed elastic stiffness per unit length along a flexible portion of the tower length:
46 n_secs_k_distr: number of points at which distributed stiffness per unit length is specified (-)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
37 38 39 40 41 42 43 44 45
: z_distr_k [row array of size n_added_m_pts; section locations wrt the flexible tower base over which distributed stiffness is specified] (m)

Dear Lingling,

You can do as you proposed, but you don’t need 1-m resolution of sections if you don’t want to. BModes will assume a linear variation between sections.

Yes, I used ADAMS to derive mode shapes during the OC3 project; this was before the cability was added to BModes.

Best regards,

Dear Jason,

Really appreciate! I still have two questions.

  1. I have tried the DS_Monopile.bmi which is modified by me from CS_monopile.bmi (please see attach and please change .xls to .bmi) with CS_monopile_tower_secs.dat which is downloaded from wind.nrel.gov/public/jjonkman/BModes/
    But, obviously DS tower section data is different from CS tower section data. Do you have the DS tower section data file or maybe the tower input file for FAST?

  2. Do you know how to come up with the distributed spring stiffness in BModes from LPile outputs? Did they do a linear curve fit since in BModes the spring stiffness is set to be linear distributed between the nodes?

Best regards,
DS_Monopile.xls (7.73 KB)

Dear Lingling,

I haven’t reviewed your DS_Monopile.bmi file in great detail, but from a quick glance it looks OK. The distributed tower data from my FAST model of the NREL 5-MW turbine atop a monopile with distributed springs (DS) is found below:

---------------------- DISTRIBUTED TOWER PROPERTIES ----------------------------
HtFract TMassDen TwFAStif TwSSStif TwGJStif TwEAStif TwFAIner TwSSIner TwFAcgOf TwSScgOf
(-) (kg/m) (Nm^2) (Nm^2) (Nm^2) (N) (kg m) (kg m) (m) (m)
0.00000 9517.14 1037.13E9 1037.13E9 798.098E9 235.129E9 41979.2 41979.2 0.0 0.0
0.45961 9517.14 1037.13E9 1037.13E9 798.098E9 235.129E9 41979.2 41979.2 0.0 0.0
0.45962 4306.51 474.49E9 474.49E9 365.133E9 106.396E9 19205.6 19205.6 0.0 0.0
0.51366 4030.44 413.08E9 413.08E9 317.878E9 99.576E9 16720.0 16720.0 0.0 0.0
0.56770 3763.45 357.83E9 357.83E9 275.356E9 92.979E9 14483.4 14483.4 0.0 0.0
0.62173 3505.52 308.30E9 308.30E9 237.242E9 86.607E9 12478.7 12478.7 0.0 0.0
0.67577 3256.66 264.08E9 264.08E9 203.220E9 80.459E9 10689.2 10689.2 0.0 0.0
0.72981 3016.86 224.80E9 224.80E9 172.987E9 74.534E9 9098.9 9098.9 0.0 0.0
0.78385 2786.13 190.06E9 190.06E9 146.252E9 68.834E9 7692.7 7692.7 0.0 0.0
0.83789 2564.46 159.49E9 159.49E9 122.735E9 63.357E9 6455.7 6455.7 0.0 0.0
0.89192 2351.87 132.77E9 132.77E9 102.167E9 58.105E9 5373.9 5373.9 0.0 0.0
0.94596 2148.34 109.54E9 109.54E9 84.291E9 53.077E9 4433.6 4433.6 0.0 0.0
1.00000 1953.87 89.49E9 89.49E9 68.863E9 48.272E9 3622.1 3622.1 0.0 0.0

I don’t know the format of the LPile output. BModes needs distributed springs (N/m per unit length along the tower or (N/m)/m), but if LPIle only provides lumped springs (N/m at discrete points) than you must distribute the spring force appropriately for use in BModes.

Best regards,

Dear Jason,

Really appreciate!

wind.nrel.gov/public/gbir/BModes-for-Erica/

I finally got the files from here who is the first author of NREL report - NREL/CP-500-43045. Just post here for somebody others may need.

Dear Jason,

Another point I am still concerning is the input for Platform.

Platform mass inertia 3X3 matrix (i_matrix_pform):
Platform-reference-point-referred hydrodynamic 6X6 matrix (hydro_M):
Platform-reference-point-referred hydrodynamic 6X6 stiffness matrix (hydro_K):

Are these parameters only being used for the floating structure such as TLP, Spar, Semi etc? If I am using monopile structure, can I just set all these parameters to zero. And I did that, the results are almost 99.99% close to the above report.

Really appreciate!

Regards,

Dear Lingling,

Yes, you may set BModes input parameters mass_pform, i_matrix_pform, hydro_M, hydro_K, and mooring_K all to zero when modeling a monopile with distributed springs (DS).

Best regards,

Dear Jason,

Really appreciate!

Do you still have the DS_tower.dat as well as the DS_platform.dat files available? Do you mind if you can send me a copy of them?

Besides that, I may still need your help to make sure the following points:

  1. Why would we not incorporate the Heave and Yaw motion of the platform?

  2. If we have set TwrLdMod = 2, does that mean that we cannot see the hydrodynamic effect (by Morison’s equation) and soil effect at the same time if we run time-marching simulation with FAST? But we can see both of them simultaneously by either a CS or an AF model.

Best regards,

Dear Lingling,

I’ve answered these questions in my Thu Feb 27, 2014 post in the forum topic found here: http://forums.nrel.gov/t/natural-frequency-of-nrel-ds-monotower-in-oc3/845/1.

Best regards,

Dear Jason,

I have the platform natural frequencies of surge or sway at around 8 Hz. And I have the platform natural frequencies of roll or pitch around 39 Hz. I just couldn’t understand the physical meaning of these four natural frequencies and why they are too high.

In my model, I don’t have any platform mass on the bottom of the mono-pile. Could you please explain what do these natural frequencies of platform stand for physically?

Really appreciate!

Dear Lingling,

I’m not sure how to answer your question without seeing the mode shapes/eigenvectors. The physical interpetation is found in the mode shapes/eigenvectors.

Best regards,

Dear Jason,

I was thinking the high natural frequencies of platform may be introduced by the last distributed spring at the pile bottom. And since there is no mass specified for the platform, the natural frequencies of them are quite high.

From the comparison of natural frequencies of tower reported by BModes and FAST, they are pretty close. I just want to make sure my model in FAST is sound enough.

I have attached the FAST Linearization output, FAST platform input, and BModes output in the attachment.

Really appreciate!

Best regards,
99.txt (118 KB)
SNLOffshr_Platform_Monopile_DS_60.4m_extreme.txt (6.14 KB)
0lin.txt (34.7 KB)

Daer LingLing,

The tower is cantilevered to the platform in FAST, and so, the tower will move with the platform DOFs. So, even if the platform has no mass/inertia, the tower will introduce an effective mass/inertia for the platform DOFs.

When I look at the eigensolution of the FAST linearization output, the 39 Hz modes seem to be strongly related to the platform surge/sway DOFs. While the 1st tower fore-aft and side-to-side modes (at around 0.17 Hz) seem to match well between the BModes and FAST models, the 2nd tower fore-aft and side-to-side modes do not (0.85/0.95 Hz in BModes versus 1.3-1.4 Hz in FAST), which is likely related to the blade flexibility in FAST (whereas BModes considers the rotor rigid when examining tower modes). It may help to run a FAST linearization analysis with only the platform and tower DOFs enabled in FAST, to ensure that the BModes and FAST models match better.

Best regards,

Dear Jason,

I agree with you!

Really appreciate!

Dear Jason,

I am still confused about how to obtain k (“Distributed elastic stiffness per unit length along a flexible portion of the tower length” in file of CS_Monopile.bmi), even though I have read this post. If it is the initial stiffness of the p-y curves, i.e. tangent on the curve at y = 0 m, k should be zero at the mudline for API p-y model (depth z=0), is it right? But this is different from the value in CS_Monopile.bmi. Please tell me in detail how to obtain k. Thanks in advance.

Best regards,
Kevin

Dear Kevin,

I can’t really provide details on how to calculate “k” because I have not done this myself. But I can respond to a couple points:

The value of “k” should be calculated as the tangent to the p-y curve–assuming that the force “p” is expressed as a force per unit length along the pile. However, the value need not be defined about the undisplaced configuration (y=0), but it could be defined about a preloaded state (e.g., based on the mean thrust load at a given wind speed) if that is desired.

Best regards,