[Dave442]

Hi all,

I have carried out an FEA analysis to investigate the deployment of a medical device in a blood vessel. I now wish to use CFD to investigate blood flow through this vessel. I was hoping to get some advice on how best to generate a CFD mesh of the volume enclosed by the vessel walls and the outer surfaces of the medical device. I have attached some images for clarity.

I can export the deformed vessel and medical device from my FEA package as either .STL meshes or as facetted .SAT geometries. Would it be a better approach to work on cleaning up the facetted geometries using a CAD package then using Booleans and whatnot to generate a solid model of the domain which I could then use to create a mesh? Or is it possible to generate a CFD mesh directly from .STL meshes?

I should mention that I have access to Ansys Workbench, Ansys CFX and Ansys ICEM (all v12.1). I am a complete novice as regards CFD and associated packages and, as such, any advice would be greatly appreciated.

Thanks in advance,
Dave

[PSYMN]

ICEM CFD can mesh directly to the STL model... You will be done your project before the other-universe-you even finished a fraction of the CAD work...

There is an STL Aorta tutorial, but any basic Patch Independent (Octree) Tetra tutorial will do, since it all works the same way. You can find some Tetra tutorials on youtube...

http://www.youtube.com/ansysinc#p/u/31/SdUjpjwUnew

http://www.youtube.com/ansysinc#p/u/30/C1Yw_pYaPGE

Is the STL surface complete? Or will it need repair? I am assuming that since it came out of the solver it is complete, but if you do need to repair it, ICEM CFD has many tools for that also.

[Dave442]

Hi Simon,

Thanks a lot for your reply. I hope I can avoid extra CAD work!

I can import the domain of interest into ICEM as a mesh, and from this mesh I can build an analytical topology. I have attached an image to show you. Im sure that the individual parts, i.e. the vessel surface and the surfaces making up the medical device are complete. Unfortunately, as these geometries were obtained from an FEA analysis, there is a significant amount of slight penetration between the two parts. This can be seen in the second image I have attached.

Im guessing this issue would cause a lot of problems with mesh generation. Would it be possible to clean up my model directly in ICEM? or would I be forced back to work with a CAD package?

Thanks again, I really appreciate the advice - Dave

[PSYMN]

Yes, ICEM CFD has all the repair tools you will need...

But ICEM CFD doesn't require a perfect model... How much penetration are we talking? Is it significantly more than the element size? If not, then don't even worry about it... Just put a material point in and mesh the model to see how it goes.

If you do have problems, such as the mesh not properly capturing features, you can just "draw" your own lines where you want them to be and octree tetra will follow those lines instead of the actual geometry... You could create those by intersection if that helps...

In the end, if you have a few bad elements due to poor geometry, ICEM CFD has all the mesh editing tools you would need to fix it and move on rather than go back and CAD your geometry.

Simon

[PSYMN]

Oh yea, you should probably start with "build diagnostic topology" just to chop up your STL into segments which you could then break up into various parts... Then you can do things like "intersections" much more easily...

[Dave442]

Hi Simon,

OK, thanks for the advice. The penetration between the two parts is negligible in terms of element size so maybe I will get away with it? To be honest I have very little experience using ICEM so I have a bit of work to do before I understand what you mean by "putting a material point in" and "octree tetra" etc..

From your advice though it appears that I can create the required meshes directly from an .STL file in ICEM which is exactly what I wanted to figure out. Thanks for your time and for the advice, I really appreciate it!

Dave