[SOLVED] composite_objects

[ewpostek]

Dear Colleagues,
I would like to build a composite object.
This is just a cell with nucleus.
Therefore, I build two meshes (gmsh) with cytoplasm and with the nucleus (different mat. properties). They have common nodes on their interface.

Could you give advise to build it ?

best regards,
Elek

This is just to load cytoplasm (outer mesh).
I intend enrich it with nucleus, but I stopped with assignment of the mat. properties.

<Node name=”cell1.msh” gravity=”0 -9.81 0″>
    <EulerImplicitSolver name=”cg_odesolver” />
    <CGLinearSolver name=”linear solver” iterations=”25″ tolerance=”1e-09″ threshold=”1e-09″ />
    <MeshGmshLoader name=”meshLoader” filename=”mesh_n/mesh_ctp_a.msh” />
    <TetrahedronSetTopologyContainer name=”topo” src=”@meshLoader” />
    <MechanicalObject name=”dofs” src=”@meshLoader” />
    <TetrahedronSetGeometryAlgorithms template=”Vec3d” name=”GeomAlgo” />
    <DiagonalMass name=”computed using mass density” massDensity=”1000″ />
    <TetrahedralCorotationalFEMForceField template=”Vec3d” name=”FEM” method=”large” poissonRatio=”0.3″ youngModulus=”100″ computeGlobalMatrix=”0″ />
</Node>

[Hugo]

Hi Elek,

Would you not rather use one single mesh with heterogeneous material properties?
This would be more coherent and easy to simulate.

About the xml file you wrote: be careful with the quote you use. Only the regular one (“) are recognized. Another user recently got an issue due to this.

Hope this helps.

Hugo

[ewpostek]

Hi Hugo,
Yes, it would be much better, definitely.
I simply do not see the coincidence
between ‘rphys’ variable and the
definition of the for example
<TetrahedralCorotationalFEMForceField template=”Vec3d” name=”FEM” method=”large” poissonRatio=”0.3″ youngModulus=”100″ computeGlobalMatrix=”0″ />

<TetrahedralCorotationalFEMForceField template=”Vec3d” name=”FEM” method=”large” poissonRatio=”0.3″ youngModulus=”500″ computeGlobalMatrix=”0″ />

since I should set it twice for two different physical models. This is the second tet.
rphys = 1
2 4 1 1 4 654 1708 1974 2101

and this is the 15032th tet with rphys = 2
15032 4 2 2 4 1319 1390 1320 3008

Because both kind of elements should be
in the same Mechanical Object.

best, Elek

Thank you for the hint on the (“)

[ewpostek]

Hi,
Sorry, may be it is redundant.
Would you mind to give a hint
how to build a composite object based
on one single mesh (gmsh 1.0 format) ?
(different kind of materials)

kindest regards, Elek

[Hugo]

Dear Elek,

What you need is to define the elasticity of each of the tetrahedra (if you use a tetrahedral mesh). This can be done by using the data youngModulus of the TetrahedronFEMForceField. The data youngModulus is a vector of real (E). The size of this vector is the number of tets.

So, what you can do is:

• load you Gmsh mesh file
• load the elasticity vector, that you can save in a .obj file as:

  
  v E1
  v E2
  v E3
  

• and give the elasticity vector to your TetrahedronFEMForceField

In a scene, it would look like:

<?xml version="1.0" ?>
<Node name="root" >
    ...
    <MeshGmshLoader name="GmshLoader" filename="path_to/your_mesh.msh" createSubelements="true" />
    <MeshObjLoader name="elasticityVectorFile" filename="path_to/elasticity_vector.obj" />

    ...
    <TetrahedronFEMForceField youngModulus="@elasticityVectorFile.position" />

</Node>

To see the heterogeneous elasticity, you can even use the option drawHeterogeneousTetra=”1″ in your TetrahedronFEMForceField
Let us know if it works.
Cheers

Hugo

[ewpostek]

Thank you, I will do the test on comparison
of the stiffnesses (hetero and homogenous structure).

So far, I have checked the reading of the file
and the applicability of the drawHeterogeneousTetra=”1″ to
the TetrahedronFEMForceField
and TetrahedralCorotationalFEMForceField

Both commands read the file *pos. However, the
TetrahedralCorotationalFEMForceField gives the warning:
Unused Attribute: “drawHeterogeneousTetra” with value: “1”.

Could you tell me if the vector *pos should work with
other parameters as well? Namely, Poisson’s ratio,
and for the nonlinear materials parameters like for example neohokean.
This is because I intend to have nonlinear viscous-elastic
finite strains materials in the bodies.

Cheers, Elek

[Hugo]

Hi Elek,

So, you succeeded in having heterogeneous Young’s modulus E, right ?

Other parameters are unfortunately homogeneous for the whole object.
Moreover, non-linear constitutive laws in SOFA exist but are in private repositories for now. Hopefully, it will soon be released open-source.

Best regards,

Hugo

[ewpostek]

Hi Hugo,
Yes, I did. There still is quite much work
to build the framework of the system before
adding nonlinearities.
Best regards, Elek.

[Hugo]

So Elek,

It means your post about getting heterogeneous elasticity is solved, right?
Cheers,

Hugo

[ewpostek]

Yes, it is,
cheers, Elek

[Hugo]

Nice to hear, I close it.
Best,

Hugo

[Author]

Posts