simtk.org

Hello,

I would like to perform an FSI ALE in a pipe made of 3 different materials. I can create a single fluid domain and 3 different solid domain (external wall) in order to assign to each of them a different constitutive law but how can I manage the interfaces between the 3 solid domain? What type of BCs Can I use to define the mechanical behavior of the wall rings at the interfaces? Is it the right approach to manage this kind of simulation? What do you suggest?

Thanks for your time

Valentina Ceserani

Are these multiple solid domains connected? Do they share a common boundary? If so, you don't have to create them as separate mesh entities, but only set Domain IDs for each of the domains where you would like to set different constitutive models.

Can you post a picture of the problem you are trying to solve? Please include the domains (fluid & solid) and the boundary conditions you want to set. I could not visualize the wall rings you are talking about.

Hello!
Thank you for your answer.

Before to consider 3 pipes, I would like to consider the problem step by step in order to understand how Simvascular works with different models.

I would like to replicate the example proposed in svTest --> "07-fsi/ale/03-pipe_3D".

I have defined a model composed by two pipes with a common boundary, as you can see in the figure attached.

I have defined 4 different domains:
LUMEN 1 --> fluid domain of the pipe1
WALL1 --> solid domain (wall) of pipe1
LUMEN 2 --> fluid domain of the pipe2
WALL 2--> solid domain (wall) of pipe2

I have imposed a pressure as inlet condition (LUMEN1-inlet), a projection between each fluid domain (LUMEN1-wall, LUMEN2-wall) and related internal solid wall (WALL1-internal, WALL2-internal), and a projection between the outlet surface of the fluid domain fo the pipe1 (LUMEN1-outlet) and inlet surface of fluid domain fo pipe 2 (LUMEN2-inlet). I have no imposed any constrain between the surfaces of the solid domains in contact (common boundary). Actually the simulation run.
I attached the file.inp.
Another simulation with different properties for each solid domain is running.


Could be it a kind of approach to manege this problem?

Following your suggestion, How can I define different ID for different solid domain tracts without creating differen models?

Thank you for your time.
Valentina

Hi,

I have looked at your input file, and it appears that you are trying to do FSI in the pipe3D svFSI-Tests but using different wall properties. If yes, then one projection that you missed is between WALL1-outlet and WALL2-inlet. Instead of a projection, you are setting Dirichlet BCs with 'Zero out perimeter: 1', which may lead to locally unexpected deformation.

If you only want to set different wall properties, you can create two domains splitting the wall mesh into two. You may create a Python script to (a) load the mesh, (b) identify the two domains using some criterion (e.g., z < L/2, L is the pipe length), (c) store these domain ids at the element level in an array DOMAIN_ID, and (d) export a vtu file or append this variable to the mesh file. Now load the domain file in your input as described here

https://github.com/SimVascular/svFSI/bl ... L281-#L302

Hope this helps.

-- Vijay

Thank you for the suggestion!
I successfully modified the .vtu file and added "DOMAIN_ID." I noticed that the same procedure needs to be followed to include information related to the initial values for both the fluid and the solid. I wanted to ask if, in this case as well, the information should be added by assigning a specific name to the array.


# Additional data may be provided depending on the equation solved
# # For FSI simulations, fluid domain may initialized as:
# Initial pressures file path: ./init/flow_from_rigid_walls.vtu
# Initial velocities file path: ./init/flow_from_rigid_walls.vtu

# # For prestress-based FSI, solid domain may be initialized as:
# Prestress file path: ./init/wall_prestress.vtu
# Initial displacements file path: ./init/wall_disps.vtu

Thank you for your time.

Valentina Ceserani