MDv2: Activating bending stiffness capability

[pschuenemann]

THIS PULL REQUEST IS NOT READY TO MERGE - Needs some review.

Feature or improvement description
When running a simulation with EI = 0 and another one with EI = 10000 with the same model I realized that the results for both simulations are exactly the same. Looking into the code I saw that there is one line commented out which prevents the user input EI to be assigned to the respective line. This issue is resolved here by uncomment the mentioned line.

However, even with this PR the simulation results (with and without EI) are still looking very similar (though they are not exactly the same anymore), which I wasn't expecting. Thus, I'm not shure whether there might not be other issues within the implementation of bending stiffness. Maybe there were also a specific reason why this line was commented out.
@mattEhall Could you please double-check?

Impacted areas of the software
MoorDyn v2

Test results, if applicable
Not available

[mattEhall]

Thanks Paul. It looks like there are a couple lingering issues with the bending stiffness implementation. In addition to the behaviour you mentioned I am also getting some NaNs when using rigid (cantilever) cable attachments. I'm trying to sort this out now...

[mattEhall]

@pschuenemann, I've enabled the Line bending stiffness and then fixed a number of related issues for application of the bending stiffness, especially where Lines connect to Rods, so that bending moments are properly transmitted. Hopefully this commit a8b1d09 now gives the correct behavior. I still have more testing to do, but let me know if this seems to solve things on your end.
Matt

[pschuenemann]

Thank you very much for this new commit @mattEhall .
Unfortunately, even with this update things are not completely solve on my end.

Now, when I simulate static equilibrium (i.e. still water and no platform motion) I see some weird periodic motions of the cable at the connection between two sections with different properties connected by a Rod (see this video for a visualization of the motions). If I reduce the number of segments (i.e. reduce parameter value NumSegs) those motions seem to disappear. However, if I increase the number of segments the results show strange node positions in the vicinity of the connection (see attached figure).

Please let me know, if I can help you on further elaborate on this issue.
Thanks again for your support and continous development.

.

[mattEhall]

I think I finally have these issues solved in a commit I just made: 8539e4e. Thanks @pschuenemann for all your rigour and clear descriptions in identifying these problematic behaviors. I'm impressed to see your example working with segments only 1m long. Let me know if the latest update solves everything!