[Farhang]

Hi everyone, I am going to use ANSYS-CFX and FSI-model a pulsating flow through a valve with a deformable diaphragm. In the initial step, the diaphragm is undeformed and prevent any flow through the valve (so there is no mesh after the diaphragm). However, when the inlet flow increases, the diaphragm is deformed and allows the flow to pass. For this situation, I need to have meshes after the diaphragm to model the flow through the valve. My question is how I should create the meshes for the whole domain of the problem?

[ghorrocks]

If you want to model something which is closed and opens you have a few options, each with their own advantages and disadvantages:
* Use a GGI as that handles opening and closing easily. But you need to slide the surfaces over each other and any valves have flat faces which close so this is a simplification.
* Start the simulation with the valve cracked open a tiny amount. Then the simulation just proceeds from there. This avoids the problem of the moment the valve cracks open but means you can't start from properly closed, and also mesh stretching in the valve seat can be an issue
* Similar to the previous option, but use a momentum source term to stop the flow when you define the valve is shut.
* The most accurate way of handling this is with dynamic remeshing, as then you can handle anything - as long as you can define the remesher to do it. This is a lot harder than the other approaches. Warning: I have no idea if dynamic remeshing works well with FSI.

In general people seem to like the hardest possible approach, so if you want to do dynamic remeshing have a look on the ANSYS customer webpage for tutorials on it.

[Farhang]

Originally Posted by ghorrocks

If you want to model something which is closed and opens you have a few options, each with their own advantages and disadvantages:
* Use a GGI as that handles opening and closing easily. But you need to slide the surfaces over each other and any valves have flat faces which close so this is a simplification.
* Start the simulation with the valve cracked open a tiny amount. Then the simulation just proceeds from there. This avoids the problem of the moment the valve cracks open but means you can't start from properly closed, and also mesh stretching in the valve seat can be an issue
* Similar to the previous option, but use a momentum source term to stop the flow when you define the valve is shut.
* The most accurate way of handling this is with dynamic remeshing, as then you can handle anything - as long as you can define the remesher to do it. This is a lot harder than the other approaches. Warning: I have no idea if dynamic remeshing works well with FSI.

In general people seem to like the hardest possible approach, so if you want to do dynamic remeshing have a look on the ANSYS customer webpage for tutorials on it.

Thank you Glen for your helpful suggestions. I have started a modelling similar to what you mentioned in your second comment and I am now waiting for the simulation to be finished. I am also doubting about the possibility of using "Dynamic re-meshing" with FSI. I will keep you updated if I get anything new.