6
$\begingroup$

I'm thinking a bit about the Front Tracking method used for simulation of Two phase flow with sharp interfaces. The literature tells me that the main issue is the surface mesh representation (operations + container) of the fluidic interface.

Does anyone here know about a good C++ library for dynamically (truly dynamical: free storage, removal and deletion of elements from memory, not fictitius logical removal) adaptive surface meshes?

I've seen some results used in the field of Computational Graphics where such surface meshes are reconstructed from point information that is recorded by cameras, and are evolved with very intensive topological changes... also, there are curvature based self-intersection algorithms out there used for cloth motion simulations... The knowledge seems to be out there, but there is not much happening in the OpenSource world.

Improve this question
$\endgroup$

3 Answers 3

Reset to default
1
$\begingroup$

Try VTK with netgen. I've been using them for years, they are very useful. but they do not come without issues.

Improve this answer
$\endgroup$
3
  • 2
    $\begingroup$ Could you specify what the issues with respect to this combination are? More detail = better and more useful answers! $\endgroup$
    – aeismail
    Jan 6, 2012 at 19:43
  • $\begingroup$ All of the mentioned surface meshing algorithms are inherently not robust, so you need to know how from experience to use them in what situations. If the points are evenly distributed, the characteristics are smooth, underlying geometry is simple etc. it will usually work. For more real cases, with topology fusions and breakups they will sooner or later break somewhere - that is why you have so much more papers here than reusable code. $\endgroup$
    – Dominik Lark
    Jan 7, 2012 at 8:12
  • $\begingroup$ The point is not to mesh the surface, that would not be possible in a way that conserves mass: the surface is actually a fluidic interface. The idea is to use a surface mesh and move it in a lagrangian way using the flow velocity, and then determine the density distribution in the two phase flow using the new surface mesh (fluidic interface) spatial configuration. Remeshing is out of the question, instead, the point is to fix self intersections and sharp curvature regions by adding-removing triangle mesh elements... $\endgroup$
    – tmaric
    Jan 7, 2012 at 13:28
1
$\begingroup$

Actually, I'm working on a very similar problem: two phase flow simulation through porous media from X-ray scanned rocks. If you don't mind using the Octave freeware, there is a program called 'ISO2MESH', which can obtain surface meshes from black & white stacked images. You can also produce tetrahedral meshes with this program, but its limited in the different file types that it can directly save to. If you don't mind coding a parser to reconstruct the geometry/meshes in another program, then this program might be of great use to you.

I almost forgot to mention: ISO2MESH is also a freeware! :) You can obtain it here.

Improve this answer
$\endgroup$
1
$\begingroup$

Try the A48 dynamic adaptive mesh library.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.