Principal stresses from Simvascular

[Tejas Canchi]

Hi, I want to use Simvascular to extract principal stress data for an aneurysm model that I am running. I see that TAWSS and OSI are extractable. But these are parameters on the fluid side. Is there a way to get the structural parameters as well?I am running an FSI simulation using the deformable RCR. Thanks!

[Hongzhi Lan]

Hi Tejas,

As I know, currently the solver only outputs the parameters for the fluid domain, not for the structure. But the solver may has some information about the structure stress during the computation. To output it, we need to modify the solver source code. I'll talk to one of our group member about this, then I'll let you know if and when we'll add this feature.

Thanks,
SimVascular Development Team

[Hongzhi Lan]

Hi Tejas,

After we discussed with our lab member, there is a possible way to calculate the structure stress. Here are some suggestions as below from Abhay, one of our lab members. Hope they help.

SimVascular Development Team

---------------------------------------------------------------------------------------------------
Given postsolver outputs displacements you can theoretically get principal stress from first principles. I say theoretically because I have never done stress calculations.

One plausible way to get principle stress,for linear elasticity, with caveats, is:

1 - You compute relative displacement = current displacement- reference displacement.(see remark 2 below)
2- gradient of relative displacement should give you deformation gradient
3- Once you have deformation gradient you can compute a bunch of strains- including linear strain
4- With strain and Youngs modulus one can get stress tensor.
5- From stress tensor you can compute principal stresses.

The solver does compute this stress internally as its imposed on the mesh as a traction term in CMM. Not sure if its in a form where you can just add a flag and write it to file.

Remarks:
1- Given output files are in vtk format its convenient to do above said steps using vtk filters/scripts.
2- Choice of ref. displacement is a discussion by itself. Though physiologically inaccurate many choose it to be diastole, for convenience. There are algorithms to compute the true ref. config. (see Hsu, Ming-Chen, and Yuri Bazilevs. Finite Elements in Analysis and Design 47.6 (2011): 593-599) or ways around (see Baek, S., et al. CMAME 196.31 (2007): 3070-3078.)
-------------------------------------------------------------------------------------------------

[Tejas Canchi]

Thanks for the reply Hongzhi. I wanted to know if i can use material models like Mooney Rivlin on the solid side of an FSI simulation.

Also is there any way to explain how to use this? As in, can I add separate constants for the model? "Number of Wall Properties per Node (10) 10,21 Defines the material model for the vessel wall. Option “10” (default option) means that an isotropic material model needs to be used. Option “21” means that an orthotropic material model needs to be used"

[Hongzhi Lan]

Hi Tejas,

Current Simvascular doesn't have any hyperelastic model like Mooney-Rivlin materials. It is a simple linear elastic model. In addition to thickness and E, no other wall properties can be employed in current setting.

Hongzhi