14 July 2020 at 18 h 05 min #16875SaraParticipant
- University of Bristol
Hi SOFA forum,
I have modelled breast tissue in SOFA under the assumptions of homogeneity and linearity, and would like to attach now a 1.5-2 mm thick ‘skin’ on top of my hemispherical model. I am not attempting to make an especially accurate model, but would like an approximate behaviour instead. How can I go about doing this? I will give some additional detail here:
– I am creating a gripper that will house sensors for the detection of tumorous tissue, and I think I have successfully done this. It is modelled using a TetrahedronFEMForceField, and will be the focus of my project.
– I have also meshed and uploaded to my scene a hemispherical breast approximation shape modelled using TetrahedronFEMForceField, with a poisson ratio of 0.495 and young modulus of 3.55 kPa.
– To test the force imposed by the gripper, and the stress created in the structure, I would like the gripper to scan a breast-like surface (this is where the approximate model of breast tissue is useful)
– There are collision meshes for both the gripper and the breast tissue, but when I try to scan the surface of the tissue, at some point the gripper will penetrate the tissue. I thought this could be due to the low young modulus of the tissue model, and so decided to model the skin as well, as this will act as a ‘casing’ with a significantly higher young modulus (about 1.2 MPa). It may also be due to something missing from my collision meshes, but I cannot tell.
I have tried to model the skin as a child of the ’tissue’ node, in 2D with a TriangularFEMForceField, but I am not sure this is the right approach. My main questions are:
1) Could the problem in bold above be due to an incomplete/poor collision mesh?
2) How can I create a skin with a stiffer material and attach it to the breast model?
Can I get some advice on where to start resolving these issues?
Thank you in advance,
Sara18 July 2020 at 13 h 39 min #16901SinghParticipant
I am curiously waiting for a helpful reply to this question as I am working on similar problems.21 July 2020 at 16 h 58 min #16917HugoKeymaster
- SOFA Consortium
To answer your main question, modeling the stiffness of the surface (skin or membrane) bounding an object, the good approach is to add a stiffness on a surface model mapped on the volumetric mechanical model.
An example of this is given in: examples/Components/topology/Tetra2TriangleTopologicalMapping.scn.
The cylinder has a constitutive law for its volumetric part, and a stiffness is added on the surface of this cylinder. It would then mimic the stiffness of a skin or similar.
Regarding your collision issue where the gripper penetrate the tissue, I would need to see the whole scene. It could be numerical settings.
I hope this helps.
Hugo23 July 2020 at 12 h 46 min #16946SaraParticipant
- University of Bristol
Thanks for your reply. Yes, I have been actually having trouble setting the alarm distance, contact distance and time step. So far, I have managed by trial and error, but as you can imagine this is very time consuming and still gives me a slow simulation with lag as well as occasional contact failure. Each time I make an adjustment to the CAD file of the gripper, I have to again keep changing these three variables randomly until the gripper can safely make contact with the breast tissue model.
Any advice you can give me regarding how I can best choose these variables would be very helpful. I have supplied meshes, controller and main script for my SOFA scene in the file below, in the case that numerical settings are not the main issue here:
My best regards,
- You must be logged in to reply to this topic.