SimScale CAE Forum

How to run a simulation with multiple geometries?

Hey everyone!

So for starters, I am a neurosurgeon and have very limited prior experience in computer simulations. I have worked in three FEA projects before, but never used the simulation engine myself, an engineer was helping us do the research.

After finding out about simscale I thought it would be great to simulate some other hypotheses in brain surgery through these kind of simulations and wanted to start a new project.

I will link the project below, and explain what I am trying to do, It would be great if anyone could help me with this!

https://www.simscale.com/projects/siyarbahadir/brain_ventricles/

So there is three geometries in this project: brain, brainstem and ventricles.
Ventricles act like a pool that stores cerebrospinal fluid (CSF) in it. The pressure inside the ventricles is usually around 90-150 mmH2O. But there are times that it increases, and increases in CSF pressure can be fatal due to compressive effect on brain and brainstem.

So what I would like to simulate is the effect of different CSF pressures in ventricles on these two structures, I ran a CFD analysis with just a ventricle and its showing me the fluid flow, but what I would like to find out about is how much force is applied to the ventricle walls and surrounding structures.

I couldnt find a way to use these three models in one simulation, can anyone help me with this?

Thanks a lot!

Hi siyarbahadir,

as long as the CSF is the only part containing the fluid it should be the only part that’s being simulated. Therefore i see no problem in your current Simulation Setup.

In regards to the force acting on the neighboring parts oft the CSF you can use your Area Average result control item for the surfaces of interest. With that results the force is calculated as

F=P*A

F=Force
P = Pressure
A = Area

To see the deformation on the Brain and brainstem, a FEM analysis would be necessary.

Also a few tips:

Please reduce the fines of your mesh. 8.Mio Cells are quite much and therefore result in long simulation times.
When you reduced the cell count you can also run your simulation for more iterations. (currently you only simulated for 15 iterations ( It should be at least 500 for a convergent simulation.)

I hope that helps with your questions.

Best regards Sebastian

Hi SBlock!

Thank you very much for your valuable contribution. I was able to get much better results.

I reduced the finenes of my mesh and now calculations are complete in minutes. I was really starting to think waiting 6 hours for a simulation was probably norm in CFD analysis.

I also calculated Area Average Results for different group of faces and it seems what I get is data of how much pressure is applied to a particular face, do I need to write a short script to calculate the force applied to each face, and is there a way for me to create a “force distribution map” around the ventricle walls and somehow translate that into the FEA analysis and calculate displacements according to that force distribution map?

Hi siyarbahadir,
great to hear that i could help. CFD can easily can run for multiple hours. That’s why you will always have to make compromises in your mesh accuracy.

When you are really interested in the accuracy of your mesh, you can create a mesh independence study.

With that you would increase the fines of your mesh step by step. At one point your results shouldn’t differ anymore and your results aren’t depending on your mesh anymore.

For applying the results of the CFD to FEA the current way is to create an Area Average report and take the pressure results and apply it to the according faces. By doing that for each face you will get a “pressure map” which you will be able to assign to a pressure boundary condition in your FEA analysis.

A tool that can help you with the set up are Topological Entity Sets,

With that you won’t have to select each face individually.

I hope that answers your questions.

Best regards Sebastian

1 Like