[cgma-dev] Meshquite and iGeom

[Tim Tautges]

Hi Dan,
   If you're interested in implementing an algorithm for restoring geometry-mesh association, I'd be very interested in 
putting that in Lasso.  I've always had the opinion it would be best to do this topologically, starting with the 
geometric vertices, then the geometric curves, etc, traversing mesh adjacencies to find connected vertices as candidates 
(once you associated all the vertices).

- tim

On 01/17/2012 09:02 AM, Iulian Grindeanu wrote:
> Hi Dan,
> What is your configuration? Did you build cgm with OpenCascade or with Cubit/Acis?
> I think it is OpenCascade.
> So are you trying to associate all mesh vertices with the closest surface geometry?
> Can you send the code you are using to test this?
>
> iBase_REGION corresponds to a volume entity in iGeom or to a 3D solid element in iMesh (3D topological entity)
>
>  From what I see, the geometry and mesh are pretty close to each other. How are you deciding what nodes are on the
> boundary? Are you using Skinner in moab?
>
> Iulian
> ------------------------------------------------------------------------------------------------------------------------
>
>     Hi,
>     I'm still trying to test the relation between mesh and geometric model. But there is some problem, when I try to use
>     iGeom (iGeom_getEntClosestPt) to compute
>     the distance from mesh coordinates and geom surfaces.
>     There 2 point which should be on geometric surface, but the computed distance are wrong:
>     10990.6 2573.73 0 (computed closest:11202.6 2486.35 0.000607243)
>     10990.6 2573.73 3000 (-->11202.6 2486.36 3000)
>     There 're no problem by other points (in ori_mesh.vtk).
>     I wonder what happened by the computing. And how exactly does iGeom_getEntClosestPt work?
>     Best Regards
>     Le, Dan
>
>     PS: what's the meaning of iBase_REGION ?
>
>     Am 16.12.2011 06:39, schrieb Iulian Grindeanu:
>
>
>             Mesquite will modify position of the free nodes, to get a better quality; "constrained nodes" can be
>             constrained on geometry entity, so they can move only along geometry. This is an advanced use of Mesquite,
>             and you should understand what you can do with it.
>
>             Sets of nodes can be constrained to move along geometry; 1d sets for geometric edges, 2d sets for faces
>             (while keeping the boundary nodes on specific boundary edges too).
>             This setup may have to be done "by hand", including generation of sets. How did you create the hexa mesh?
>             The nodes that are on the boundary may have to be put in sets, for all your edges, faces, etc.
>             You have a unit issue with your model; I think that in theory you can "squeeze" that vtk file to fit that
>             wall geometry, but I do not think that this is the proper way to use mesquite. This is why I suggested to
>             look at meshkit examples (only one mesquite test involves geometry, and it is too simple)
>
>             One would be test_onetooneswept in meshkit/test/algs.
>
>             Iulian
>
>         I was wrong with the suggestion to try test_onetooneswept in meshkit/test/algs.
>         Until this evening, it did not work with OpenCascade;
>         It should work now, if you update MeshKit
>         Also, in Lasso, test_entset will work if you launch it with
>         test_ent brick.stp brick.h5m (update Lasso too, to get those files)
>
>         Iulian
>
>
>
>

-- 
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (tautges at mcs.anl.gov)      (telecommuting from UW-Madison)
  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
             fax: (608) 263-4499      Madison, WI 53706