Hi Ganesh,
yes I confirm that their are extra nodes, and some of them are coming from the source topological geometry. These nodes are the easiest to eliminate, as they are not interconnected with the element array. Even there may be others that are.
21.12.2015, 13:25, "Ganesh Diwan" <[email protected]>:
So you got for one spot more than one node?-No - I did not check if at one physical location if I have more than one node. Sorry but I am not sure what you mean by detection method.
How did you detect the there are double nodes, then? With paraview? Or just view the .msh file? Did you notice it in the gmsh gui?
-which was the export format?I am exporting it in standard Gmsh .msh format.
ok, thanks, .msh is standard for Gmsh. I see.
you say you want the true mesh, the physical mesh. So "physical" means that it mixture of body, boundary and pure mesh information.When I created a mesh, without doing Geometry -> Physical Groups ->Add->Surface, I only had a 'Elementry entity' of the surface. So, when I exported the mesh, I could see that there are nodes which are in $Nodes section but are not part of the finite element connectivity in $Elements section. In fact these extra nodes were created at the exact physical locations where I had specified the points to define a B-spline bounding line. Nothing wrong wit
pardon, are you really sure of this? Is there a mathematical evidence? How do you know, any implementation of following processing code is proof to it? No?
You are not sure even about it, if it is a bug, then one should report it, therefore a bug demo script were not bad.
Admittedly I am not much time at this list, and maybe have not studied thoroughly its history/bug list.
But it seems that questions as these are somewhat difficult. Maybe boringly asked to much with enduring completion state?
Pardon: do you know how I deal with it? I have a python script that filters all double nodes out. And do you why I do this? Because they told me to do so in the solvers forum. Imagine!
From my point of knowledge it is a bug, and needs to be dumped in the bug tracker or should be already there. When I go for it, I leave a message to this thread - and yes - would be glad if other did.
Regards
this except I do not need these nodes for my FE program. Specifying a physical group as above however enables the export of only those nodes that "belong" to the physical surface - exactly what I need.Hope this helpsGanesh
On Sat, Dec 19, 2015 at 10:29 PM, T. Feldhaus <[email protected]> wrote:15.12.2015, 23:14, "Ganesh Diwan" <[email protected]>:,GaneshThanks in advance for helpDear GMshI create a 2D mesh from a surface with a B-spline as its bounding line. However, when its mesh is exported, all the points originally used to create the B-spline also get exported as nodes of the mesh. Is there a way to export only the true mesh i.e. nodes and elements that belong to the physical surface alone and not B-spline.Hi,-would like to understand, what you are saying: So you got for one spot more than one node?--which is the detection method and-which was the export format?-you say you want the true mesh, the physical mesh. So "physical" means that it mixture of body, boundary and pure mesh information. But you find something over more inside, topological points from the original geometry as doubles?tia_______________________________________________
gmsh mailing list
[email protected]
http://www.geuz.org/mailman/listinfo/gmsh
_______________________________________________ gmsh mailing list [email protected] http://www.geuz.org/mailman/listinfo/gmsh
