[Owen]

Dear Hugo,

Thank you for your reply!

This new mechanism could be extremely useful for dynamics randomisation. I am going to give it a try. Thank you and all the other developers’ great effort! I will close the post now.

Best wishes,
Owen

[Owen]

Hi @hugo

I am sorry to bother you. I guess you are very busy at the moment but I would like to know if there is any recent update on updating parameters in FEMForceFields like ‘TetrahedronFEMForceField.youngModulus’. Can you share some info and insights?

Thank you,
Owen

[Owen]

Hi,

I have resolved the problem by installing a Ubuntu 18.04 virtual machine. Now MOR is running smoothly. Thus I am going to close this post now. I want to go a big thanks to Olivier for returning to Python2 is the right thing to do to solve the problem.

Best,
Owen

[Owen]

Hi @olivier-goury,

Yes, I agree that going back to Python2 could solve the error. The only reason I choose to use Python3 is that I found it difficult to install PyQt4 with my current version of Qt5, while PyQt5 only supports Python3.

If I understand it correctly, PyQt is for the GUI so it is essentially optional for the plugin. I am wondering if there is a no-GUI version of MOR that could let me avoid the usage of PyQt?

Best,
Owen

[Owen]

Update: the syntax error seems to be fixed after I changed originalScene = r’$ORIGINALSCENE’ to originalScene = r”$ORIGINALSCENE” in the phase1_snapshots template.

However, this ‘fix’ might not be correct as I got a new error:

========= STDERR-LOG============
CUDA: 1 device(s) found.
CUDA:  0 : "Quadro T2000", 3112/3903 MB, 128 cores at 1.785 GHz, revision 7.5, timeout enabled
CUDA: Using device 0 : "Quadro T2000"
[ERROR]   [PythonScript] IOError: [Errno 2] No such file or directory
  File "phase1_snapshots.py", line 27, in <module>
    originalScene = imp.load_source(originalScene.split(slash)[-1], originalScene)

[ERROR]   [SofaPython] Script (file:phase1_snapshots) import error
[ERROR]   [SofaPython] scene script load error.

I have emailed Olivier for his opinion. Will update the post with his valuable suggestions.

Best,
Owen

[Owen]

Hi @Hugo,

Sadly, I am experiencing the same problem described in this post. As you mentioned earlier ‘This is due to the fact that the constraint resolution has no CUDA version (CudaVec instantiation) in the open-source version of SOFA. However, it does exist in some private plugins.’, do you refer to the SOFA Multicore-GPU plugin? If so, I think I will need to apply for access in order to continue what I am doing…

Best,
Owen

[Owen]

Hi @hugo

I am trying to use the Ogden model for my simulation as well. I am very pleased to see the TetrahedronHyperelasticityFEMForceField contains Ogden, but I am struggling to set the parameters.

As you mentioned in your previous reply, in Ogden.h:

Real k0=param.parameterArray[0];
Real mu1=param.parameterArray[1];
Real alpha1=param.parameterArray[2];

Considering the equation of the strain energy density in the Odgen model, does it mean N is assumed to be 1 for the Ogden model on SOFA? I am wondering how to set the parameters if we have 3*mu and 3*alpha in an Odgen model where N=3.

Many thanks,
Owen

[Owen]

Hi @hugo,

I feel kinda stupid to waste your time on fixing this tiny error… However, I have one more question to ask about BilateralInteractionConstraint if you don’t mind. I found this page explaining the Args: https://sofacomponents.readthedocs.io/en/latest/_autosummary/_autosummary/sofacomponents.ConstraintSet.BilateralInteractionConstraint.html

But I still don’t understand what first_point and second_point represent. For example, in my case, I want to combine two objects so I created two points on the body of each object. What would my first_point and second_point be?

Best,
Owen

[Owen]

Hi @secretdevil

Thank you for your reply!

I have corrected the case of “Constraint” as Hugo suggested, and then I don’t have this error anymore. I am sorry that my code is a bit messy… Even though I don’t fully understand how the templates work, I think the mechanical state was created in the children of rootNote. Like in Line 126 pneu2.createObject('MechanicalObject', name='tetras', template='Vec3d', showObject='0', showObjectScale='1') etc.

I guess for now I will move on since the error disappears. If I encounter similar problems again, I will examine my code and find where Vec3d is missing.

Best wishes,
Owen

[Owen]

Hi JN,

I tried the first suggestion and it works! You have really made my day. Thanks a million!

Best,
Owen

[Owen]

Hi JN,

Thank you for your reply.

Yes, the full output seems a bit too long. I just redirected it to a text file. Please find the file in the link: https://drive.google.com/open?id=1sfmqP_7E-r6_ziejZGYV-lICtGFjVBaY

Many thanks,
Owen

[Owen]

Hi Hugo,

Thank you for your quick reply.

You solved my problems! Everything works perfectly in the console. I just realised that there is so much potential in using python script.

Thank you!

Regards,
Owen

[Owen]

Thank you Hugo. I am gonna close the topic now.

Owen

[Owen]

Hi Hugo,

Sure, after removing the borders, I used cellSize=’2′, facetSize=’2′, cellRatio=’2′, facetApproximation=’0.15′. However, they only give me an acceptable mesh because I am running SOFA on my laptop which is not very powerful.

I guess I will start with BilateralInteractionConstraint. Is there any example code I can take a look?

Many thanks,
Owen

[Owen]

Hi Hugo,

Thank you very much for your reply!

You are absolutely right about the meshes. The thin borders caused my problem. I have managed to simulate the actuator after I removed them. So thank you!

My next step is to simulate an assembly of two of this actuator. However, I couldn’t find a tutorial on how to define the connection between two bodies on SOFA. Could you give me some hints?

Cheers!
Owen

[Owen]

The link to the STL files was lost somehow… I hope this one works

https://drive.google.com/drive/folders/1U1tdDmSaEXBaW1JNx1Qad-OtlfkTK8ES?usp=sharing

[Author]

Posts