Forum Message

 

 

We have moved the forum to https://forum.ngsolve.org . This is an archived version of the topics until 05/05/23. All the topics were moved to the new forum and conversations can be continued there. This forum is just kept as legacy to not invalidate old links. If you want to continue a conversation just look for the topic in the new forum.

Notice

The forum is in read only mode.

Second order meshing of 2D domains (with multiple subdomains)

More
3 years 10 months ago #2887 by bhaveshshrimali
Second order meshing of 2D domains (with multiple subdomains) was created by bhaveshshrimali
Hi everyone,
I was trying to generate a second order mesh for a domain comprising of a circle inside square. But was facing an issue. Here is the MWE with a very coarse mesh to delineate the problem:
Code:
from ngsolve import * from netgen.geom2d import SplineGeometry periodic = SplineGeometry() pnts = [ (0,0), (1,0), (1,1), (0,1) ] pnums = [periodic.AppendPoint(*p) for p in pnts] top = periodic.Append( ["line", pnums[2], pnums[3]], bc="outer") right = periodic.Append ( ["line", pnums[1], pnums[2]], bc="periodic") btom = periodic.Append( ["line", pnums[1], pnums[0]], leftdomain=0, rightdomain=1, copy=top, bc="outer") left = periodic.Append( ["line", pnums[0], pnums[3]], leftdomain=0, rightdomain=1, copy=right, bc="periodic") # adding the circle periodic.AddCircle((0.5, 0.5), 0.3, leftdomain=0, rightdomain=1, bc = "inner") msh = periodic.GenerateMesh(maxh=5) msh.SecondOrder()

This, however, generates a mesh where the mid-points of the edges along the circle aren't properly placed. I am guessing it has to do with how I am defining the geometry



Any suggestions on what could be going wrong.

Thank you
Attachments:
More
3 years 10 months ago - 3 years 10 months ago #2888 by bhaveshshrimali
Replied by bhaveshshrimali on topic Second order meshing of 2D domains (with multiple subdomains)
[attachment=undefined]paraview_Ordered.png[/attachment]I think I confused `triplot` of the corresponding linear mesh with the actual mesh. However I noticed something. The ordering of vertices is slightly different than what one would expect in Paraview (vertices followed by edge midpoints in the same order as the vertices)

For instance:
Code:
points = [] cell_arr = [] for pt in msh.Points(): points.append([*pt]) points = np.array(points) for elem in msh.Elements2D(): cell_arr.append([int(str(i)) for i in elem.points]) points = points[:, :2] cell_arr = np.array(cell_arr, int)-1
when fed to paraview gives:

Attachment not found



whereas adjusting the last 2 points as
Code:
cell_arr = cell_arr[:, [0,1,2,5,3,4]] #the last three points are ordered different than what one would expect

gives the desired ordering

[attachment=undefined]paraview_Ordered.png[/attachment]
Last edit: 3 years 10 months ago by bhaveshshrimali.
Time to create page: 0.140 seconds

Login