build_obb error witn .exo file created by Cubit 14.0

[Wei Li]

Dear DAGMC Developers,

Thank you for developing this wonderful tool and I am just learning to work with dagmc and OpenMC.

I created a simple cylinder body in Cubit 14.0 and exported it as the "exo" format. I used "mbconvert" of MOAB to successfully convert it into "h5m" file. But when running openmc, error showed up:

----------------------------------------------------------------------------------------------------------------

 Reading settings XML file...

 Reading cross sections XML file...

 Reading materials XML file...

 Reading DAGMC geometry...

Loading file dagmc.h5m

Initializing the GeomQueryTool...

Using faceting tolerance: 0

Building OBB Tree...

[0]MOAB ERROR: --------------------- Error Message ------------------------------------

[0]MOAB ERROR: Failed to join the obb trees!

[0]MOAB ERROR: construct_obb_tree() line 496 in src/GeomTopoTool.cpp

[0]MOAB ERROR: --------------------- Error Message ------------------------------------

[0]MOAB ERROR: Failed to construct obb tree for volume!

[0]MOAB ERROR: construct_obb_trees() line 600 in src/GeomTopoTool.cpp

[0]MOAB ERROR: --------------------- Error Message ------------------------------------

[0]MOAB ERROR: Failed to build obb trees!

[0]MOAB ERROR: setup_obbs() line 172 in DagMC.cpp

[0]MOAB ERROR: --------------------- Error Message ------------------------------------

[0]MOAB ERROR: Failed to setup the OBBs!

[0]MOAB ERROR: init_OBBTree() line 204 in DagMC.cpp

[0]MOAB ERROR: load_dagmc_geometry() line 186 in dagmc.cpp

 WARNING: No graveyard volume found in the DagMC model.This may result in lost

          particles and rapid simulation failure.

 ERROR: Could not find a root universe.  Make sure there are no circular

        dependencies in the geometry.

--------------------------------------------------------------------------

MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD

with errorcode -1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.

You may or may not see output from other processes, depending on

exactly when Open MPI kills them.

----------------------------------------------------------------------------------------------------------------------------------------------------------------------

The journal file of Cubit I used is as follows:

create cylinder height 1 radius 0.4

surface 1 3 2 scheme trimesh

mesh surface 1 3 2

set duplicate block elements off

block 1 surface 1 2 3

undo group begin

set large exodus file off

export mesh "/home/wli/pin.exo" overwrite

undo group end

I think this is caused by the build_obb function of DAGMC, as the command "build_obb dagmc.h5m" gave the same error.

Is there anything I should pay attention to when generating Cubit file for MOAB or DAGMC? 

Thank you very much in advance.

Best,

Wei

[Andrew Davis]

Hi Wei

Typically we use Cubit/Trelis as a pre-processing tool, and for which we have written a dedicated plugin/extension that allows one to produce DAGMC geometries directly as in the website instructions https://svalinn.github.io/DAGMC/usersguide/trelis_basics.html. So in the background information you will see that we produce (effectively) a surface mesh based representation of the geometry via a process called faceting as opposed to surface meshing as you've described here. Now we can have a conversation about perhaps supporting that method of producing geometry (as you've described) but currently, I dont know exactly how moab would treat the topology of the geometry that you have exported in exodus format.

So, I recommend - if you can - get a newer version of Cubit say 15 onwards and use the plugins that we describe in the documentation, you will have more success that way. 

Thanks

Andy

--
You received this message because you are subscribed to the Google Groups "DAGMC Users and Collaborators" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dagmc-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dagmc-users/e51c2336-fb45-4a33-8a61-93f35e1c55d3%40googlegroups.com.

[Wei Li]

Hi Andy,

Thank you for your prompt reply. I may ask my advisor if I can get a higher version of Cubit.

Thanks,

Best,

Wei

在 2020年1月30日星期四 UTC-5下午4:03:20，Andrew Davis写道：

To unsubscribe from this group and stop receiving emails from it, send an email to dagmc...@googlegroups.com.