OrcaFlex - What is 'nodal period'

[DCockey]

Originally Posted by daiquiri

DCockey, looks like we don't speak the same language here... So, I'll use graphics to express myself.
What I am saying is this: (see the attached picture). Considering the uncertainities in every single factor involved in the mooring design, I think that any more complex model than one in the pic would be an overkill.
So if ones' analysis of mooring lines involve elements with 0.01 s (100 Hz) natural periods, I really have to agree with Gonzo's last phrase in the post #20.

Your illustration may be close to how OrcaFlex models the lines.

Based on Floating's description in his post #7 his problem originated with trying to model a bridle so that it have a smooth cantenary curve AND the use of the explicit solver. I'm assuming the bridle is a relatively short part of the mooring line system. Whether the elasticity of the bridle needs to be modeled depends on what results are needed and on the general approach to the model. Everyone appears to agree that it's not relevant to the general motion of the buoy or lines.

If the peak loads need to be resolved then elasticity may need to be considered. When a system is modeled as rigid then peak loads can be infinite in some cases. Elasticity spreads the peak out.

[daiquiri]

Quite oddly, there's one thing we have never considered in this thread:

- the physical parameters used by Floating for modeling the mooring line could be wrong, resulting in such short time periods.

[DCockey]

I suspect it's related to the period of axial vibrations of the line, not transverse vibrations. Using spring and lumped mass system then the natural period of an axial vibration would be:
T = 2 * Pi * sqrt (mass / stiffness)
Line of length L, area A, density Rho, modulus E:
mass = Rho * L * A
stiffness = A * E / L
Putting these into the equation for the period
T = 2 * Pi * sqrt [(Rho * L * A) / (A * E / L)] = 2 * Pi * L * sqrt (Rho / E)
So for axial vibrations the diameter and area don't affect the natural frequency and period. (NOTE: The above equation over-estimates the axial vibration period for a cable with distributed mass. It is correct for the approximation of a lumped mass at the end of a spring.)

I don't have the parameters for the line he used so I'm unable to calculate a period.

[DCockey]

Originally Posted by daiquiri

I've found some data for Spectra ropes here: http://www.machovec.com/rope/spectra_12strand.htm

Don't know what diameter are your lines, but consider an example of a 40 mm rope. Its lineic density (from the data sheet in that webpage) is 98 kgs/100 ft = 3.2 kg/m. ...

That chart has a mistake. The column heading should be kg /100 m, not kg / 100 ft. For the 40 mm rope it shows 65.8 lb / 100 ft which is 98 kg / 100 m or 0.98 kg/m

[daiquiri]

Originally Posted by DCockey

I don't have the parameters for the line he used so I'm unable to calculate a period.

Then let's improvise.
Let's take this rope as example, since it gives the elongation vs. tensile load: http://www.novabraid.com/pdfs/P20101-2%20Novablue.pdf
Considering just the linear part of the elongation curve (for tensions higher than 10% of Tmax) you get, with some math and for a rope with 40 mm diameter:
E = 2.74 GPa
rho = 992 kg/mc

Hence:

T = 3.8E-3 * L

For a 30 meters long line, it would give a period of 0.11 s, or 8.8 Hz natural frequency.
It would take line less than 2.5 meters long to get 100 Hz natural frequency.

Originally Posted by DCockey

That chart has a mistake. The column heading should be kg /100 m, not kg / 100 ft. For the 40 mm rope it shows 65.8 lb / 100 ft which is 98 kg / 100 m or 0.98 kg/m

Correct observation. The pdf in the above link looks more realistic imho.
Redoing the calcs in the posts #3 and #4 with this new data, the relationship between pre-tension, frequency and length for a transversally oscillating rope becomes:

Tension = 5 f^2 L^2

Maximum tension is 345 kN, so to get 100 Hz a rope shorter than 2.6 meters is necessary, almost identical value as seen before.

[DCockey]

So an element 1 to 2 m long would have a natural frequency above 100 Hz and a period less than 0.01 sec. That's in accordance with what Floating reported.

Again, what matters for the time intervals of an explicit method is the period of the highest frequency mode of the discrete system as modeled, and in this case that corresponds to the frequency of a particular single element.

[DCockey]

The reasons for the short time step is the use of the explicit solver in conjunction with the short elements considered necessary to model the bridle part of the mooring system. An implicit solver, as Floating recommends in post #11, uses considerably longer time steps, and may even adjust the time step during the calculation; shorter when needed, longer when feasible.

Perhaps the amount of resolution used in modeling the bridle isn't really required for what Floating needs, but I won't even guess without considerably more knowledge of his needs and the system he is modeling. It sounds like once he switch to the implicit solver and determined an appropriate set of parameters he was able to get the solutions he needed.

[floating]

Thanks to all for the helpful discussion. Some background, this is a mooring in moderate water depth (40m). The bridle connects via a tendon to a surface float then a mooring line. The overall mooring system has a reasonable scope for the depth. Each mooring component must be broken up into short line segments in the model, even though it is composed of a single line in reality. I am modeling this mooring's behavior in a 100-year storm, and am now using implicit rather than explicit integration.

OrcaFlex tech support has kindly looked at my model, and indeed I did make poor choices on line properties which caused some of the spurious high-frequency motions (though I am not clear why): the overall mooring system did not have enough drag and inertia, and I should have divided the bridle and some other components into many more segments. My 'solution' of modeling the bridle (source of highest-frequency behavior) as Links instead of Lines wouldn't have solved these problems.

Once I digest these changes, I'll report back if I learn anything more of interest. Again thanks to all.

[floating]

Originally Posted by DCockey

I suspect it's related to the period of axial vibrations of the line, not transverse vibrations.

Yes it is. And if I had used OrcaFlex's Modal Analysis tool, I would have seen that right away: it finds the highest-frequency mode for each model component and specifies whether it is axial or transverse.

[DCockey]

Thanks for reporting back. I was afraid we may have scared you off.