Meshing low thickness region

  • 206 Views
  • Last Post 21 August 2018
  • Topic Is Solved
miguwhynot posted this 19 July 2018

Hello, I'm currently doing a project whose geometry is a small cube of air that contains a balloon made of a small plastic film of 2 mm thickness. The inside of the balloon is helium and I'm not being able to correctly mesh the geometry of the film.

I've tried to use different meshing operation and method such as sizing, Hex and lowering the Max Size Face. But I keep getting different warning which result in the film not being mesh. Do you know what should I do in my case?

PS: I attached some images about balloon geometry and the more common error. I' using Ansys 19, Student version.

Error: Patch-conforming tetrahedral mesh failed because of duplicated nodes in boundary mesh.

Attached Files

Order By: Standard | Newest | Votes
rwoolhou posted this 20 July 2018

Which solver are you planning on using? It may be you can model the balloon as a thin surface which may simplify things. The other issue may be if the balloon touches the cube as this will create a very poor mesh at that point, and it might be this that's preventing the thin volume from meshing. 

Try pasting images into the text as the attached files are unreadable here! This will also enable the community to give you some more pointers. 

Cheers,
Rob

 

 

miguwhynot posted this 20 July 2018

Thanks for your comment. I'm planning on using FLUENT for heat tranfer study purposes. The ballon does not touch cube faces if that's what u meant, but it does touch the air supposed to be cointained in it. Here i give you some images:

I hope the images are clear enough, in any case thanks again for your help. I'll be glad to five any more info about my project.

 

Cheers,

peteroznewman posted this 20 July 2018

Hello,

You could apply two vertical planes of symmetry through the center of the balloon and air cube at the Y-axis, and end up with a 1/4 model. 

That will give the mesher much easier geometry to mesh and you a much smaller model (or a more detailed mesh) to solve.

I assume you subtracted the balloon and helium volumes from the air volume cube and set Shared Topology to Share.

If you try this and are still having issues, then it may be worth using Rob's suggestion of changing the balloon to a surface model instead of a solid. The surface will be easier to mesh.

Regards,

Peter

miguwhynot posted this 20 August 2018

Hello,

Sorry for answering so late. I have tried using symmetry plane, but that didn't solve the problem.

Here's a better explanation to the problem:

I'm trying to model the heat transfer of a high alttitude balloon, during the ascension. I'm planning on using FLUENT since I've read some work related to this topic that used this program(also is the oneI'm most familiar with). 

I need three volumes, one for the air flow(outside the ballon), one for the balloon skin and also one for the hellium inside the ballon.

Here's a more complete screen on what the geommetry looks like:

 

I've been able to make it work with higher skin thickness. But in reality the skin is only 0.2 mm, I think thats where the problem is coming from. Sizing the pesh in the skin part could maybe be a solution, but then again the number of elements would be enormous.

I also tried to simplify the skin as a surface, but that resulted in a super strange mesh interpretation by FLUENT.

rwoolhou posted this 20 August 2018

OK, I think your best approach is to use a thin wall in Fluent. Essentially we can ignore the geometric thickness of the balloon but account for the heat transfer in the solver. This way you just need two volumes: inside & outside. 

 

  • Liked by
  • José Mantovani
  • miguwhynot
miguwhynot posted this 20 August 2018

Thank you so much for your time, I think this could very well be the solution!

OK, so here's what I did in the meshing app:

Here's a closer look on what i meant by "skin wall":

And here's how I think the wall should look in Fluent:

 

 

Does it look good? Again can't thank you so much both of you!

rwoolhou posted this 20 August 2018

That looks about right: I can't see that well because of the image resolution. When you display a cut plane in meshing there's a little pale blue tetrahedron on that panel: this shows the full cells making it easier to see what the mesh looks like. I'd also recommend reading up on y+ with respect to heat transfer and flow. 

  • Liked by
  • miguwhynot
miguwhynot posted this 20 August 2018

Can't thank you enough for this! This will really simplify the problem without adding symmetry (I would love to get a clear cut inside the balloon natural convection).

I'm afraid I don't really know what y+ is, could you elaborate?

Thank you so much!

rwoolhou posted this 21 August 2018

It's a value that's used to determine the near wall cell height. You're best off using Google for this one: we've got slides in the training courses but you need either Customer Portal access or a Learning Hub Subscription to access these. 

 

  • Liked by
  • miguwhynot
Close