I'm stuck with the problem that I've received an corrupted or badly designed model which contains a lot of plate elements which have the shape of a triangle, but are actually quadrilateral elements.
Hereby forming all kinds of false free edges in the middle of an surface area.
Any tips or tricks to solve this as fast as possible?
Solved! Go to Solution.
The most efficient way to arrive to a quality mesh is to CLEANUP the geometry, surely your surface geometry have SMALL EDGES (segments of curves with very small length) then the mesher can not make miracles.
Can you detail more your problem? I see many ways a quad can have a tria shape:
1) 2 consecutive edges aligned. If that is the case then using
a) Check > Element Quality > FEMAP Internal Angle 90° => you group all bad elements
b) mesh > Editing > diagonal split these elements => problem solved
But that wouldn't create free edges so that might not be your problem
2) quad referencing twice the same node, or 2 nodes which are in fact coincident. FEMAP will issue a warning message, and again you can group them using the element quality methods. use Jacobian > 0.95 if internal angles get confusing.
Again I don't see how this would create free edges.
The problem doesn't occur in vicinity of small edges, since they're just in a plate field (mesh size ~ 500).
It follows more like the description of @astrium_tls; somehow two nodes have collapsed while meshing, but without creating a tria. And indeed, when checking with jacobian > 0.95, they all show up.
But what might be a fast way to convert those elements to tri (there are more then 800..)
OK, still can't see why there are free edges, but let's pass.
In the case of nodes collapsed onto each other, your problem is that "Check Coincident Nodes" can't act. So you need to start by splitting!
1) create a group with bad elements, using Jacobian for example, or aspect ratio, or taper...Etc... some element quality check
2) split these bad quads diagonally (Mesh > Editing > Split) => this will necessarily create one "good" tria and one that is very slim
3) find all the very slim trias using the aspect ratio quality check => delete them!
4) now "Check Coincident Nodes" can work its magic, and voila!
Thanks, that helped a lot.
The free edges occured due to the fact that it couldn't match the apparent coincident edge, from which one contained actually '3 nodes'. This made all attached curves in the plane appearing as a free edge due to the double node...