Dear all,
I am working in the prostate edema modeling where i want to simulate the interaction of swelling prostate with the neighbor organs using finite element analysis. In my development i use the FeBio software.
I set contacts between the prostate and the neighbor organs (bladder and rectum). Initialy they are separate and while the prostate swells interaction between the organs occur. My problem is that if i constrain the bladder and rectum the prostate penetrates and if I let some degrees of freedom the prostate pushes away the two other organs.
The desired response would be the prostate pushing the two other organs in some degree while all the three organs undergo shape deformation due to the contact pressure between them. For a reason I can't figure out why I can't produce this response. Has anyone an idea why this is happening?
ps. I consider uncompressible materials for the prostate and rectum and poroelastic material for the prostate.
Thank you in advance,
Konstantinos.
Contact problem of two or more deformable bodies is highly complex, so usually there are a lot of things you have to set up or you can optionally set.
Maybe there are problems with boundary conditions, but not certainly. Let me suggest an idea.
When I applied Marc for simulation of cutting processes (large deformation and deformation rate), there was a typical problem I had to face. Bodies penetrated if displacement of nodes on their surface during an iteration step was larger (or comparable) with contact tolerance. So it may be useful for you to check that estimated (by you) displacement of nodes in contact is smaller significantly than contact tolerance. If not, then decreasing time step of iteration, or increasing contact tolerance may help you.
Be careful with contact tolerance, because if it is too large, then results may be prematured (i.e. bad), and if too small then computational cost may become very large, and detecting contact status may lead to numerical difficulties.
I will suggest to include spring elements among the organs. In this way, all the organs will move differently. You need to define the stiffness (k) of the springs by using F=kx. In additión, it will be also interesting to include gap elements within the places that you have with an initial space.
There is more information needed. Where do you constrain the neighboring organs? How? "Pushing away" is equivalent to some space "behind" the organs. Which are the properties of this space? Is it filled with another kind of deformable material? Incompressible bladder and rectum looks ok - which material model is used for the deformation of these neighbors? Uriel's idea of spring elements is promising ... however, as far as I remember there are no gap elements in FEBio. I am sorry. I have more questions then answers. 3D contact problems are very nasty. Could you provide a raw sketch of your model?
Dear all,
Thank you for your suggestions and interest. To make more clear my model answering to dr Prüfer i am attaching an image showing highlighted the constrained surfaces of the model. The organs are constraint by fixing in zero displacements at these surfaces in all the x, y, z axes. the space between of the organs is assumed material free. The material for bladder and rectum is Mooney-Rivlin.
Thank you all for your useful recommendations.
Konstantinos
Hi Konstantinos,
at first sight, your model seems to be correct. I do hope it will not be too embarrassing: Did you perform some convergence tests with "your" organs? As I mentioned before contact is really nasty. Moreover, you have complex contact surfaces.
I recommend a stepwise procedure. Perform a contact analysis rigid body with rigid body motion against the bladder, e.g., with a sphere as contactor. If you get realistic deformations, try the contact deformable sphere against the (deformable) bladder. The problem in your model is the multiple contact. You do not have very much control by mere intuition.
If you have already done such tests, please forget the previous paragraph.
Contact is lots of trial and error, mostly error.
Greetz, hp
I always start with the simplest model. Then I know where the problem is.
The simplest means the linear without contact and next with contact (it can be assumed a soft filling between bodies that there will be no negative Jacobian).
Linear materials with poisson coeficient of 0.45-0.49.
If it works well, then enter the contact. If it works, then the problem is in material models. It seems to me that that the complex material models (and to this different) can be problematic. Are You sure that sofware can calculate properly contact with such different incompressible materials ?
Set materials as linear and try ( so will be the fastest). I think that constrains are not the problem.
I'm curious about the results, as cyclist {:-|
- Badges
- Science topic