- This topic has 13 replies, 3 voices, and was last updated 4 years, 11 months ago by
.
-
Posts
-
Hello everyone,
I want to model a slack soft rope in any curved shape at the beginning. When there is no axial force, it can turn into any curved shape after a collision. But it can also stretch when it is exerted tensile force. And part of it would turn thinner when it is being clipped.
How can I do it?
Wong
Hi Wong,
I am very new in SOFA but triyng to do similar issue. I think, you question has to be answered in two Levels:
1)Definition of the physical model of your soft rope (decision for mechanical structure)
2)Computational modelling of this model.To 1) I think – in your case the most simple way to represent the behaviour is to model the structure as it is. What you describe seems to be real braided rope, it behaves exactly as you describe. Generally, in FEM You can model the single yarns as truesses, beams or solids.
How it can be done, you can read of some of my books and papers, for instance
https://www.springer.com/de/book/9783030025403
or
directly using their implementation in Texmind Braider Software. Sorry that it looks like advertisment, but I had to develop this, because I did not found anything else working for me and I had to teach this topic last decade. What I am trying now is to prepare direct Export from there to Sofa Scene, but have some troubles understanding the definitions of the beam orientation and similar issues. If you can help in this – the solution will be faster. In other case some way trough some intermediate Format should be possible
To 2) how to model this in Sofa – this should then run depending on the boundary conditions without problems after that.Hello @yordan-kyosev
I think the situation is more complicated than you think. What I want to model is something like vessels. So you have to think on the case that the rope has branches.
As far as I can see, representation of beam is still nothing more than a strings of rigid bodies. It may cause low performance having masses of rigid frames and it can never deform radially when it is being clipped.
And I cannot buy your book in my country as well.
But I am still interested in your work and I still want you to provide more information if it is still possible to help.Thanks,
WongHi @outtt
All beam approaches can not be deformed radially since it is actually the definition of the method: one direction (length) is significantly greater than the two other ones (section).
In the community, vessels are usually represented by beams because the added stiffness due to their presence is of interest. They are also sometimes modeled as static triangulated surfaces when we talk about endovascular catheter navigation.
A third model could be possible is the shell model. I know a private plugin with such an implementation exists.I hope this helps.
Best,Hugo
Hi @Wong,
I do not see the problem of using beams for the wires or yarns (not for the complete vessel)
We have done it with LS-Dyna for another “not classical” Geometry here
https://iopscience.iop.org/article/10.1088/1757-899X/406/1/012025/meta
and have similar not published results for bifurcated braids.
If the same approach would work with SOFA depends on that if beams can be represented there and if they capabilites to deal with contact between them.Hi @outtt
It is a private plugin belonging to a research team. I don’t know the plugin myself but I can gather information if needed. Then, the discussion would be pursued per email for academic or commercial licensing if it fits your needs.
However, I think before triggering this administrative step, it would be good to understand in which purpose you are modeling the blood vessels: what kind of effort / forces would you like to apply to your vessels?
Depending on this, we could choose the best numerical strategy.
Best wishes,Hugo
Hello @hugo,
Just deform flexibly like the real one, from all directions and keep the performance in persisting contact. Because the vessels are located in the gap of the soft tissues.
Wong
OK
Just a last question to know what level of accuracy would be needed: which surgery / procedure would you like to model?Hugo
Hello @hugo,
Just want to drag or block the artery/vessels and clip the aneurysm. So the level of accuracy should be fairly high.
Wong
Ok then shell element makes really sense.
Can you poke me again mid-December when I’ll get back from holidays to ask about this plugin?Best
Hugo
Dear @outtt
My apologies for this long silence! The start of 2020 is intense!
Let me first ask some further info. Do you really need the blood vessel to have an accurate mechanical behavior or are you mostly interesting in the interaction/grasping effect ?
In case you want to get a ‘realistic interaction’ only, the use of triangular FE model + bending springs might be sufficient.
In case you really need an advanced mechanical surface model (shell), I do have a good news: the shell plugin will soon be made open source. Beforehand, I need to clean it and go through some administrative steps. But it will be available freely!
Best wishes,
Hugo
- You must be logged in to reply to this topic.
