<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Hello every one,<br>
I attach some new files which can help in defining the problem.<br>
In the file test.geo there is the definition of a circular domain.
<br>
The goal of the test is to generate a uniform anisotropic mesh
with characteristic sizes h1=0.1 and h2=0.5. I tried to set the
following orientations: +-45 deg. <br>
In the first case the metric tensor's components are: m11=52,
m12=48, m22=52. <br>
In the second case the metric tensor's components are: m11=52,
m12=-48, m22=52. <br>
<br>
If I insert the previous values of the metric tensor's components
in the MathEvalAniso (which is commented in the test.geo file)
everything works fine, for both orientation. <br>
<br>
However, if I try to define the metric field by using a background
mesh I get the required anisotropic mesh only for the orientation<br>
+45 deg (m11=52, m12=48, m22=52) while for the orientation -45 deg
(m11=52, m12=-48, m22=52) I get an ISOtropic mesh.<br>
The metric file is defined by using the Tensor Triangular object
and refers to a background isotropic mesh (background.msh).<br>
I attach the metric files that I have generated for the two cases:
metric_file_plus45.mtr (which works) and metric_file_minus45.mtr
(which doesn't work).<br>
<br>
It is strange that everything works fine when I directly define
the metric by using the MathEvalAniso field while it does not work
as expected when I use the metric file.<br>
Maybe I'm doing an error in the syntax of the metric file but it
seems to me that it's correct: for each triangle there are 9
values to define the coordinates of the triangle's vertices and 27
values for the definition of the 3x3 tensor in the 3 nodes of the
triangle.<br>
I use the following command to generate the mesh when using the
metric file: gmsh -2 test.geo -algo bamg -bgm metric_file.mtr<br>
<br>
All suggestions are welcome.<br>
Thank you very much for your help<br>
<br>
Andrea<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
On 16/10/2015 11:48, Andrea Ferrero wrote:<br>
</div>
<blockquote cite="mid:5620C7D4.3010605@polito.it" type="cite">Dear
Gmsh users,
<br>
I am trying to generate anisotropic meshes by the BAMG tool
included in Gmsh.
<br>
I can't get anisotropic elements when the elements are oriented
to certain directions.
<br>
<br>
For example, I generated an isotropic background mesh on a
circular domain (radius=2) which I used to define a metric file.
In each node of the this background mesh I define the metric
tensor with the same parameters for all the nodes. In particular,
I set h1=0.1, h2=0.5 and theta=pi/4. The obtained metric field
shows the following parameters: m11=52, m22=52, m12=48.
<br>
If I use these parameters I get the required anisotropic mesh
shown in the image aniso.png.
<br>
<br>
However, if I set theta=-pi/4 or theta=3*pi/4 I get the isotropic
mesh reported in the image iso.png. I would expect to get an
anisotropic mesh also with these parameters but with the elements
aligned to a direction perpendicular to the previous case.
<br>
In this second case the parameters of the metric field are m11=52,
m22=52, m12=-48.
<br>
Why do I get an isotropic mesh when theta=-pi/4?
<br>
<br>
I use the following command to generate the mesh with Gmsh 2.10.1:
<br>
gmsh -2 test.geo -algo bamg -bgm metric_file.mtr -anisoMax 100
<br>
<br>
Thank you very much
<br>
<br>
Andrea
<br>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
gmsh mailing list
<a class="moz-txt-link-abbreviated" href="mailto:gmsh@geuz.org">gmsh@geuz.org</a>
<a class="moz-txt-link-freetext" href="http://www.geuz.org/mailman/listinfo/gmsh">http://www.geuz.org/mailman/listinfo/gmsh</a>
</pre>
</blockquote>
<br>
</body>
</html>