Issues with VTU File Conversion and Utilization in SOFA pneumatic simulations.

[ndolphin-github]

Hi,

I am currently working on a project involving a soft, bellows-like robot and am utilizing code from the SoftRobots.Inverse examples in SOFA. My goal is to convert mesh files of my robot and run them with the following example code, which loads the Stanford bunny in pneumatic actuation:

bunny = rootNode.addChild('bunny')
    bunny.addObject('EulerImplicitSolver')
    bunny.addObject('SparseLDLSolver', template="CompressedRowSparseMatrixd")
    load = bunny.addObject('MeshVTKLoader', name='loader', filename=path + 'Hollow_Stanford_Bunny.vtu')
    bunny.addObject('MeshTopology', src='@loader', name='container')
    bunny.addObject('MechanicalObject')
    bunny.addObject('UniformMass', totalMass=0.5)
    bunny.addObject('TetrahedronSetTopologyContainer', position=load.position.getLinkPath(), tetras=load.tetras.getLinkPath(), name='container')
    bunny.addObject('TetrahedronFEMForceField', poissonRatio=0.45, youngModulus=60000)
    bunny.addObject('BoxROI', name='boxROI', box=[0, 0, 0, 30, 30, 5], drawBoxes=True)
    bunny.addObject('RestShapeSpringsForceField', points=bunny.boxROI.indices.linkpath, stiffness=1e12)
    bunny.addObject('LinearSolverConstraintCorrection')

However, I am encountering a couple of challenges:

VTU File Conversion and Parsing:
I am attempting to create a VTU file that is compatible and readable with the above code, similar to how 'Hollow_Stanford_Bunny.vtu' is utilized. Despite my efforts to format the VTU file similarly to the Stanford Bunny example, I am consistently encountering XML parsing errors. My current approach involves converting a VTK file to a VTU file using Python libraries. Is there a specific method or guideline for generating VTU mesh files that are compatible with SOFA?

Errors with Tetrahedron FEM Force Field:
I have noticed frequent errors when applying the Tetrahedron FEM (TetrahedronFEMForceField) tools to my designed mesh files, even after converting them into a volumetric mesh using the Gmsh software. Are there particular considerations or limitations regarding mesh configuration when applying the FEMForceField in SOFA?

Any insights or advice on these matters would be greatly appreciated. Thank you in advance for your time and assistance.

[EulalieCoevoet]

Hello,

If you have a .vtk file, there's no need to convert it. As the name of the component suggests, the "MeshVTKLoader" should load .vtk files. Do you get errors when you do that?

[EulalieCoevoet]

Could you describe these errors ?
You just need a volume mesh for the FEM and a surface mesh for the cavity, the format of the files has no importance.

[ndolphin-github]

The cavity file does not shows error but the main robot in VTU or VTK shows error reading in tetrahedron mesh error as well as the FEM Force Field. Are there any condition I have to consider when placing the cavity file and the main body such as their relative positions or other properties?

[ERROR] [TetrahedronFEMForceField(TetrahedronFEMForceField)] object must have a tetrahedric topology. The component is inactivated. To remove this error message please add a tetrahedric topology component to your scene.
model = rootNode.addChild('model')

[EulalieCoevoet]

The placement of the cavity surface mesh is important of course, it should correspond to the cavity of the volume mesh.

Do the loader and topology component show errors ? If yes can you share the errors ?

From the runSofa GUI can you double click on the loader component and check that there are tetras ?
Same for the topology component.

Can you share maybe the mesh and / or your scene?

[ndolphin-github]

@EulalieCoevoet Sorry for the delayed response. I've discovered that my mesh is ill-conditioned in certain segments, and as you mentioned, whether it's VTU or VTK doesn't make a difference. Thank you for your guidance, regardless.