[MOAB-dev] Geodesic mesh representation

[Chris Eldred]

This makes sense. I guess my concern was that I had coordinates in 3D
and topology was 2D.

So the mesh will have a dimension of 3 but the highest-dimensionality
elements in it will be faces (2D)- and MOAB can deal with this?

And yes my mesh looks basically identical to the MPAS mesh you attached.

-Chris

On Wed, Nov 6, 2013 at 5:32 PM, Iulian Grindeanu <iulian at mcs.anl.gov> wrote:
> you maybe can follow this example:
>
> https://bitbucket.org/fathomteam/moab/src/bee2e77e416976393a63a221a814d85285403a7e/examples/old/FileRead.cpp?at=master
>
> It is a 2d mesh, but it reads from 2 files, one with coordinates position
> and one with connectivity for triangles.
>
> You can create a mesh with just create_vertex() and create_element() methods
> on moab::Interface
>
> ________________________________
>
> I have the mesh in a flat text file format- one file with vertex
> coordinates, one file that lists the connectivity of faces to vertices
> (ie for each face, what are the vertices that "bound" it).
>
> Topologically, the mesh is 2D since the highest dimensionality element
> is a face (bounded by edges). However, the mesh represents a
> discretization of the surface of a sphere, so vertex coordinates are
> 3D (x,y,z). Since I am on the surface of a sphere of unit radius, I
> could always go to a 2D coordinate representation in terms of angles
> and just store the 3D coordinates as Tag values. Then everything would
> be 2D (although the coordinates would be weird).
>
> It is unclear to me how to define a 3D mesh from this, since there
> really are no cells (elements of dimension = 3) in the mesh.
>
> On Wed, Nov 6, 2013 at 4:44 PM, Iulian Grindeanu <iulian at mcs.anl.gov> wrote:
>>
>>
>> ________________________________
>>
>> Dear Moab-Dev,
>>
>> I have a geodesic mesh (like MPAS or GCRM) and I was wondering what
>> the best way to represent the mesh in MOAB was. It is topologically 2D
>> but it is embedded in a 3D manifold (so vertex coordinates are x,y,z).
>> I was thinking that I would treat the mesh as 2D, feed in "bogus"
>> vertex coordinates and just use tags over the mesh elements to keep
>> track of the actual mesh coordinates. This way, I keep the 2D
>> topological structure (which is more important for my application
>> anyways). Is this a good approach?
>>
>> Thanks,
>> Chris Eldred
>>
>> --
>> Chris Eldred
>> DOE Computational Science Graduate Fellow
>> Graduate Student, Atmospheric Science, Colorado State University
>> B.S. Applied Computational Physics, Carnegie Mellon University, 2009
>> chris.eldred at gmail.com / celdred at atmos.colostate.edu
>>
>> Probably I don't understand what you need :(
>>
>> What kind of format do you have ? netcdf? Is the mesh on a sphere /
>> surface?
>>
>> The vertex coordinates are retrieved with "get_coords" methods, and set
>> with
>> set_coords methods, or use "coords_iterate" type methods in readUtil
>> interface.
>>
>> Why not keep the full 3d mesh? what is so special about the 2d topological
>> structure?
>> Is it structured in any way? Can you use the MPAS reader or not?
>>
>> Iulian
>>
>
>
>
> --
> Chris Eldred
> DOE Computational Science Graduate Fellow
> Graduate Student, Atmospheric Science, Colorado State University
> B.S. Applied Computational Physics, Carnegie Mellon University, 2009
> chris.eldred at gmail.com / celdred at atmos.colostate.edu
>
>



-- 
Chris Eldred
DOE Computational Science Graduate Fellow
Graduate Student, Atmospheric Science, Colorado State University
B.S. Applied Computational Physics, Carnegie Mellon University, 2009
chris.eldred at gmail.com / celdred at atmos.colostate.edu