writing xmf files for unstrucutred mesh

[Kris Kuhlman]

I am writing both the "optional" ELEMENTS and VERTICES sections in my unstructured mesh (along with the required CELLS and CONNECTIONS blocks), and I have the PRINT_PRIMAL_GRID flag in the OUTPUT block, but I still don't get any xmf files.

I have a python script that creates them, so I have a sort of work-around, but it would be nice to have this work correctly.

This came up last year, and there was some discussion about how to possibly fix it, but I am not sure exactly what happened.

[Karra, Satish]

Is this for all variables or just velocity?

---

From: pflotra...@googlegroups.com <pflotra...@googlegroups.com> on behalf of Kris Kuhlman <kristophe...@gmail.com>
Sent: Wednesday, August 2, 2017 9:03:15 AM
To: pflotran-users
Subject: [pflotran-users: 4343] writing xmf files for unstrucutred mesh

 

[Kris Kuhlman]

I don't get any xmf files. I currently write my own xmf file that points to the mesh info to one h5 file (which I write myself), and use the h5 file written by pflotran for the variables (except I have to add velocity data from the vtk output, as indicated in my other post).

Kris

[Karra, Satish]

Can you send me your input deck?

---

From: Kris Kuhlman <kristophe...@gmail.com>
Sent: Wednesday, August 2, 2017 9:19:35 AM
To: pflotran-users
Cc: Karra, Satish
Subject: Re: [pflotran-users: 4343] writing xmf files for unstrucutred mesh

 

[Kris Kuhlman]

In putting together the test case, I see I am now getting pflotran to write xmf files. Sorry for the mailing list noise. 

I am not certain what I did wrong before or what I changed, but I see it now works.

I still don't get the velocity data written in the xmf file or in the h5 file (my other post). 

[Karra, Satish]

I am not sure if the velocity data has been implemented for explicit format.

 

—Satish

 

[Kris Kuhlman]

I get velocity data in vtk format for explicit unstructured, just not in h5 format. 

Kris

[Kris Kuhlman]

just as a sort of post-script, I will say the xmf files generated by pflotran are not working for me. 

I am now remembering this whole discussion last year about the output being for the dual mesh, rather than the block elements used by pflotran.

I have a pieced-together system that works for me, so this is not an issue (except maybe document better what pflotran is doing/expecting regarding output for explicit meshes).

Kris