[MOAB-dev] Question about TRIs and TETs

[Iulian Grindeanu]

----- Original Message -----


Hi there, 

I am currently working on a Finite Element problem involving discrete fracture at the University of Glasgow. 

For the last few months I have been working on a crack insertion algorithm. However, I am wondering if I could ask for some assistance on a particular problem which I will try and explain as best as I can. 

I test the tractions crossing the boundary between 2 elements and if these tractions are greater than some value I tag the face to be split. The first this I do when splitting the faces in split the mid-nodes of the edges of the face (I am using higher order elements). 

I am not sure here what you mean. maybe "The first thing I do when splitting the faces is split the mid-nodes of the edges of the face (I am using higher order elements)" ? 
Anyway, the midnodes are associated to the edges. When you split, do you mean "duplicate" the midnode? 
I thought you are tagging faces, not edges. 
It is still not clear, the "crack" will start at the midnode? If it starts at a vertex, moab works fine. If you start at a midnode, moab will not create automatically a new face (see ticket http://trac.mcs.anl.gov/projects/ITAPS/ticket/263, it has your name on it :)) 


<blockquote>
When I am doing this I use MOAB to find all adjacent 2D faces to the 1D edge. I then create a new node and this node is inserted into the new connectivity for the 2D face over the old, or original, node. This is done for every face that is connected to the edge. This obviously introduces the problem that there can be 2 2D faces adjacent to the 1D edge that can lie on the same 3D tetrahedron. 
</blockquote>
The midnode is on the edge, isn't it? 
Do you loop over faces or over edges? Anyway, the edge is adjacent to the face, and tetra, and edge is adjacent to face, too. 
The issue moab is having is that the midnode is "ignored" when deciding adjacency; this is what the ticket is about; a triangle ABCDEF (DEF mid nodes on edges), is considered adjacent to an edge ABM (AB end nodes, M the middle node), which might not suit your needs. 



<blockquote>
So I was wondering if there is anyway to use MOAB to have both of these new nodes in the connectivity of the 3D tetrahedron so that using functions I can get the connectivity of the 2D faces? 

</blockquote>
Why 2 nodes? One edge has one mid node. 

<blockquote>

Regards 

Graeme 

PS. If that sounds a bit muddled or confused don't hesitate to get in touch with questions before responding. 



</blockquote>
yes, it is unclear 

<blockquote>
-------------------- 

Graeme A. Edwards 

PhD Student - Materials & Mechanics Research Group 

School of Engineering 

University of Glasgow 
Rankine Building 
Oakfield Avenue 
G12 8LT 

</blockquote>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20120627/77ff157e/attachment.html>