CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   Requesting Help from Expert Meshers-> Biomedical Applications (http://www.cfd-online.com/Forums/ansys-meshing/83389-requesting-help-expert-meshers-biomedical-applications.html)

EphemeralMemory December 27, 2010 00:23

Requesting Help from Expert Meshers-> Biomedical Applications
 
3 Attachment(s)
Thanks for looking at this post, you're a big help.

I am quie adept at handling meshes now, but I have a few questions. In short, I'm not too sure how to fix the mesh I have... Autofixing tools are largely impractical in what I'm trying to do. I'm working with Ansys ICEM CFD.

I'm currently working with an STL of a spine, with the CSF space in the middle (giving it somewhat an annulus shape), and intermittent holes distribuited along the sides of the spine, from the csf space outwards (so its a "broken" annulus). I attached a couple of pictures of the spine that I am working with, so you can get a general idea. There is a uncovered cylinder-like shape running through the middle of the spine, and I am trying to mesh the area in-between the inner area and outer area (so, the case of an annulis, the outer area minus the inner area).

This STL was registered from MRI data, and because MRI images have relatively poor resolution, it cannot be represented perfectly.

My problem is, I cannot fix it adequately enough so that FLUENT will accept it. I have tried manual fixing tools, as well as autofixing tools, but I have fixed it yet. I have not been working with ANSYS ICEM CFD for long, however, so I thought I'd try to pool the knowledge of the experts. Any ideas?

BTW: I have also uploaded a image of a common problem area: every problem associated with multiple edges, overlapping elements etc seems to arise with these areas that look differently. Upon closer inspection, I have found that these areas apparently violate the annulus rule, where the outer and inner radii are identical (which is physiologically impossible).

Thanks for your help

PSYMN December 27, 2010 10:48

Cutplane.
 
Go into Edit Mesh => Quality and generate a quality histogram...

Click on a lower bar of the histogram (maybe everything below 0.1) to display them on the screen. Right click on the histogram and add to subset.

Pick a section and run a cut-plane thru (across the spline) so we can see the annulus thru a section with bad quality.

Turn on the volume elements.

Right click on shells and choose color by quality.

Take a screen shot of a bad area cross section and maybe one of a good area cross section... Once I have a better idea of the problem, I can suggest a fix.

But in the end, garbage in means garbage out. If your geometry is wrong, your mesh can't be much better.

Simon

EphemeralMemory December 27, 2010 15:54

Thanks for the advice, I will try it out once I get into lab tomorrow. I am aware of the fact that trash in = trash out... I am trying to do everything I can to make what I have work.

The person in charge of registering the MRI to STL is a bit of a idiot in my opinion, so I had my doubts from the beginning.

Once I try it out and check the quality, I can get back to you with a more specific identification of the problem.

Thanks again for your advice

EphemeralMemory January 6, 2011 15:55

Hey, sorry for reopening this topic:

I managed to eliminate all of the problem elements: when I run the check mesh option it reads 0 errors.

One problem I still have however is the quality of the volumetric mesh: every time I volumetrically mesh it to export to Fluent, it says the mesh is bad.

Have any ideas on why this is? If it is useful to you, I can upload the project file I created for you to view.

Thanks for your help

PSYMN January 8, 2011 15:23

Hands on help...
 
Re: http://www.cfd-online.com/Forums/ans...tml#post289522


The "mesh is bad" doesn't give me enough info to help... You can usually find all the problems before you get to Fluent, but I can take a quick look and see what is going on.

You can send me the files at Simon.Pereira@ANSYS.com

Please include your name (Real and CFD-Online) and maybe a little about what you are doing and why (I am just interested).

Also include a link back to this thread so I can connect the dots back when I am ready to respond.

Best regards,

Simon

tcarrigan August 3, 2011 11:44

Post Geometry
 
I know the post is old, but do you think you could post the geometry?


All times are GMT -4. The time now is 02:17.