Problem in the interactions between the cable constraint and TetrahedronHyperelasticityFEMForceField

[pasqualeferr94]

Hello,
I am doing a simulation about a multi-material cable-driven soft finger. My problem is:

• When I impose my two materials be linear elastic -----> everything works fine.
• When I impose one of my two materials (where the cable is attached), hyperelastic ----> the solver does crazy things.

Hence, i suspect there is a problem between the cable constraint in the soft robot plugin and the hyperelastic force field (TetrahedronHyperelasticityFEMForceField).

in the simulation i use : EulerImplicitSolver with SparseLDLSolver and GenericConstraintCorrection
my sofa version is: 23.12 on Ubuntu 22.04 release.

I will attach my scene in Python3 and also my design, maybe someone can check what i did wrong (or if it is an internal error of Sofa). the name of the file is cable_finger_ogden.py.

Many thanks in advance
cable_finger.zip

[hugtalbot]

Hi @pasqualeferr94

Using an hyperelastic material (non-linear relationship between stress and strain) makes the system non-linear. However, in SOFA only one Newton iteration is performed by default. With highly non-linear problems, several iterations of Newton-Raphson are usually compulsory to converge. I think your simulation scenario falls into this case.

Could you try using the VariationalSymplecticSolver which allows several Newton step?
With @bakpaul we were thinking implementing a generic Newton-Raphson algorithm for dynamic cases (Newton-Raphson is already available in the StaticSolver)

Hope this helps

[pasqualeferr94]

Thanks @hugtalbot for the suggestion! But the problem is not completely resolved. Now i used the Variational solver instead of the EulerImplicit, the unwanted oscillations are removed, but now seems that the cable is bending in the opposite direction (furthermore it doesn´t increase the bending increasing the force), this doesn´t happen with Eulerimplicit. I will pass to you the updated scene, to see what is the problem now. I remerber to you, the scene to check is cable_finger_Ogden.py. Many thanks
cable_finger.zip

[hugtalbot]

Hi @pasqualeferr94

The simulation being really slow I struggle to see the behavior you describe.
I think this one additional motivation to implement then generic Newton-Raphson algorithm mentioned above @bakpaul

Can you move forward using linear elasticity temporarily @pasqualeferr94 ?

[pasqualeferr94]

Hello @hugtalbot,
Thanks anyway, can you update me when the general Newton Raphson method Will be implemented?

[hugtalbot]

definitely! poke us whenever you need an update if I did not get back to you 🙌

[github-actions[bot]]

⚠️ ⚠️ ⚠️
@pasqualeferr94
Feedback has been given to you by the project reviewers, however we have not received a response from you. Without further news in the coming weeks, this discussion will be automatically closed in order to keep this forum clean and fresh 🌱 Thank you for your understanding

[github-actions[bot]]

⚠️ ⚠️ ⚠️
@pasqualeferr94
In accordance with our forum management policy, the last reply is more than 4 months old and is therefore closed. Our objective is to keep the forum up to date and offer the best support experience.

Please feel free to reopen it if the topic is still active by providing us with an update. Please feel free to open a new thread at any time - we'll be happy to help where and when we can.