smesh/doc/salome/gui/SMESH/input/constructing_meshes.doc

/*!

\page constructing_meshes_page Constructing meshes

\n Construction of a mesh on some geometry consists of:
<ul>
  <li> \ref create_mesh_anchor "Creating of a mesh object"</li>
  <li> \ref evaluate_anchor "Evaluating mesh size"</li>
  <li> \ref preview_anchor "Previewing the mesh"</li>
  <li> \ref submesh_order_anchor "Changing submesh priority"</li>
  <li> \ref compute_anchor "Computing the mesh"</li>
</ul>
Mesh can be \ref use_existing_anchor "computed using your own meshing algorithms" 
written in Python.


\anchor create_mesh_anchor
<h2>Creation of a mesh object</h2>
<em>To construct a mesh:</em>
<ol>
  <li>Select a geometrical object for meshing.</li>
  <li>In the \b Mesh menu select <b>Create Mesh</b> or click <em>"Create
      Mesh"</em> button in the toolbar. 

    \image html image32.png
    <em>"Create Mesh" button</em>

    The following dialog box will appear: 

    \image html createmesh-inv.png
    <br>
  </li>
  <li>Apply \subpage basic_meshing_algos_page "meshing algorithms" and
    \subpage about_hypo_page "hypotheses" which will be used at computation of
    this mesh.

    "Create mesh" dialog box contains several tab pages titled \b 3D,
    \b 2D, \b 1D and \b 0D. The title of each page reflects the
    dimension of the CAD model (geometry) the algorithms listed on
    this page affect to. For example, \b 3D page lists algorithms
    that affect 3D geometrical objects (solids).

    \note
    - Some page(s) can be disabled - if the source geometrical
    object does not include shapes (sub-shapes) of the corresponding
    dimension(s). For example, if input object is a geometrical face,
    \b 3D page is disabled.
    - Some algorithms affect on geometry of several dimensions,
    i.e. "1D-2D" or "1D-2D-3D". If such algorithm is selected by the
    user, dialog box pages related to the corresponding lower level
    dimensions are disabled.
    - \b 0D page does not refer to the 0D elements, but to 0D
    geometry (vertices). Mesh module does not provide algorithms that
    produce 0D elements. Currently \b 0D page provides only one
    algorithm "Segments around vertex" that allows specyfing required
    size of mesh edges about some selected vertex(vertices).

    For example, you need to mesh a 3D object.

    First, type the name for your mesh in the \b Name box, by default,
    it is "Mesh_1". Then select the geometrical object you wish to
    mesh in the Object Browser and click "Select" button near \b Geometry
    field (if name of the object not yet appeared in \b Geometry field).

    \image html image120.png
    <em>"Select" button</em>

    Now you can define 3D Algorithm and 3D Hypotheses, which will be
    applied to solids of your geometrical object. Click the <em>"Add
      Hypothesis"</em>  button to add a hypothesis.

    \image html image121.png
    <em>"Add Hypothesis" button</em>

    Click the <em>"Edit Hypothesis"</em> button to change values for the
    current hypothesis.

    \image html image122.png
    <em>"Edit Hypothesis" button</em>

    Most standard 2D and 3D algorithms can work without hypotheses
    using some default parameters. The use of additional hypotheses
    is optional (i.e. you may leave "None" in this box).

    Proceed in the same way with 2D and 1D Algorithms and Hypotheses that
    will be used to mesh faces and edges of your geometry. (Note
    that any object has edges, even if their existence is not
    apparent, for example, a sphere has 4 edges). Note that the
    choice of hypotheses and of an algorithm of lower dimension depends on
    the algorithm. 

    Some algorithms generate mesh of several dimensions while others, of
    only one dimension. In the latter case there must be one Algorithm and zero or several
    Hypotheses for each dimension of your object, otherwise you will
    not get any mesh at all. Of course, if you wish to mesh a face,
    which is a 2D object, you don't need to define 3D Algorithm and
    Hypotheses.

    In the <b>Object Browser</b> the structure of the new mesh will be
    displayed as follows:

      \image html image88.jpg

    It contains:
    <ul>
      <li>a reference to the geometrical object on the basis of
        which the mesh has been constructed;</li> 
      <li><b>Applied hypotheses</b> folder containing the references
        to the hypotheses applied at the construction of the mesh;</li>
      <li><b>Applied algorithms</b> folder containing the references
        to the algorithms applied at the construction of the mesh.</li> 
    </ul>

    There is an alternative way to assign Algorithms and Hypotheses by
    clicking <b>Assign a set of hypotheses</b> button and selecting among
    pre-defined sets of hypotheses. In addition to the standard
    sets of hypotheses, it is possible to create custom sets by editing
    CustomMeshers.xml file located in the home directory. CustomMeshers.xml
    file must describe sets of hypotheses in the
    same way as ${SMESH_ROOT_DIR}/share/salome/resources/smesh/StdMeshers.xml 
    file does (sets of hypotheses are enclosed between <hypotheses-set-group>
      tags).

      \image html hypo_sets.png
      List of sets of hypotheses: <em>[custom]</em>
      automatically added to the sets defined by the user
  </li>
</ol>

Consider trying a sample script for construction of a mesh from our 
\ref tui_creating_meshes_page "TUI Scripts" section.

\anchor evaluate_anchor
<h2>Evaluating mesh size</h2>

After the mesh object is created and all hypotheses are assigned and
before \ref compute_anchor "Compute" operation, it is possible to
calculate the eventual mesh size. For this, select the mesh in
the <b>Object Browser</b> and from the \b Mesh menu select \b
Evaluate. The result of evaluation will be displayed in the following
information box: 

\image html mesh_evaluation_succeed.png

\anchor preview_anchor
<h2>Previewing the mesh</h2>

Before \ref compute_anchor "the mesh computation", it is also possible
to see the mesh preview.

For this, select the mesh in the Object Browser. From the \b Mesh menu
select \b Preview or click "Preview" button in the toolbar or activate
"Preview" item from the pop-up menu.

\image html mesh_precompute.png
<em>"Preview" button</em>

Select <b>1D mesh</b> or <b>2D mesh</b> preview mode in the Preview dialog. 

\image html preview_mesh_1D.png "1D mesh preview shows nodes computed on geometry edges"
<br>
\image html preview_mesh_2D.png "2D mesh preview shows edge mesh elements, computed on geometry faces"

<b>Compute</b> button computes the whole mesh.

When the Preview dialog is closed, the question about the storage of temporarily
created mesh elements appers:

\image html preview_tmp_data.png

These elements can be kept in the mesh.


\anchor submesh_order_anchor
<h2>Changing submesh priority</h2>

If the mesh contains concurrent \ref constructing_submeshes_page "submeshes", 
it is possible to change the priority of their computation, i.e. to
change the priority of applying algorithms to the shared sub-shapes of
the Mesh shape.

<em>To change submesh priority:</em>

Choose "Change submesh priority" from the Mesh menu or a pop-up
menu. The opened dialog shows a list of submeshes in the order of
their priority. 

There is an example of submesh order modifications of the Mesh created on a Box
shape. The main Mesh object:
<ul>
  <li><i>1D</i> <b>Wire discretisation</b> with <b>Number of Segments</b>=20</li>
  <li><i>2D</i> <b>Triangle (Mefisto)</b> with Hypothesis<b>Max Element Area</b>
  </li>
</ul>
The first submesh object <b>Submesh_1</b> created on <b>Face_1</b> is:
<ul>
  <li><i>1D</i> <b>Wire discretisation</b> with <b>Number of Segments</b>=4</li>
  <li><i>2D</i> <b>Triangle (Mefisto)</b> with Hypothesis<b>MaxElementArea</b>=1200</li>
</ul>
The second submesh object <b>Submesh_2</b> created on <b>Face_2</b> is:
<ul>
  <li><i>1D</i> <b>Wire discretisation</b> with <b>Number of Segments</b>=8</li>
  <li><i>2D</i> <b>Triangle (Mefisto)</b> with Hypothesis<b>MaxElementArea</b>=1200</li>
</ul>

And the last submesh object <b>Submesh_3</b> created on <b>Face_3</b> is:
<ul>
  <li><i>1D</i> <b>Wire discretisation</b> with <b>Number of Segments</b>=12</li>
  <li><i>2D</i> <b>Triangle (Mefisto)</b> with Hypothesis<b>MaxElementArea</b>=1200</li>
</ul>

The sub-meshes become concurrent if they share sub-shapes that can be
meshed with different algorithms (or different hypothesises). In the
example, we have three submeshes with concurrent algorithms, because
they have different hypotheses.

The first mesh computation is made with:
<center>
\image html mesh_order_123.png
<em>"Mesh order SubMesh_1, SubMesh_2, SubMesh_3"</em></center>
<center>
\image html mesh_order_123_res.png
<em>"Result mesh with order SubMesh_1, SubMesh_2, SubMesh_3 "</em></center>

The next mesh computation is made with:
<center>
\image html mesh_order_213.png
<em>"Mesh order SubMesh_2, SubMesh_1, SubMesh_3"</em></center>
<center>
\image html mesh_order_213_res.png
<em>"Result mesh with order SubMesh_2, SubMesh_1, SubMesh_3 "</em></center>

And the last mesh computation is made with:
<center>
\image html mesh_order_321.png
<em>"Mesh order SubMesh_3, SubMesh_2, SubMesh_1"</em></center>
<center>\image html mesh_order_321_res.png
<em>"Result mesh with order SubMesh_3, SubMesh_2, SubMesh_1 "</em></center>

As we can see, each mesh computation has a different number of result
elements and a different mesh discretisation on the shared edges (the edges 
that are shared between <b>Face_1</b>, <b>Face_2</b> and <b>Face_3</b>)

Additionally, submesh priority (the order of applied algorithms) can
be modified not only in a separate dialog box, but also in
the <b>Preview</b>. This helps to preview different mesh results,
modifying the order of submeshes. 
<center>
\image html mesh_order_preview.png
<em>"Preview with submesh priority list box"</em></center>

If there are no concurrent submeshes under the Mesh object, the user
will see the following information.
<center>
\image html mesh_order_no_concurrent.png
<em>"No concurrent submeshes detected"</em></center>


\anchor compute_anchor
<h2>Computing the mesh</h2>

It is equally possible to skip  \ref evaluate_anchor "the Evaluation"
and \ref preview_anchor "the Preview" and to \b Compute the mesh after
the hypotheses are assigned. For this, select your mesh in
the <b>Object Browser</b>. From the \b Mesh menu select \b Compute or
click "Compute" button of the toolbar.

\image html image28.png
<em>"Compute" button</em>

The Mesh Computation information box appears.

\image html meshcomputationsucceed.png

If the mesh computation failed, the information about the cause of the
failure is provided.

\image html meshcomputationfail.png

After you select the error, <b>Show Sub-shape</b> button allows
visualizing in magenta the geometrical entity that causes it.

\image html failed_computation.png 
<em>3D algorithm failed to compute mesh on a box shown using <b>Show
    Sub-shape</b> button</em>


\note Mesh Computation Information box does not appear if you set
"Mesh computation/Show a computation result notification" preference 
to the "Never" value. This option gives the possibility to control mesh
computation reporting. There are the following possibilities: always
show information box, only if an error occurs or never. 
By default, the information box is always shown after mesh computation operation.

<b>Publish Sub-shape</b> button publishes the sub-shape, whose meshing
failed, in GEOM component as a child of the mesh geometry, which
allows analyzing the problem geometry and creating a submesh on it in
order to locally tune hypotheses.

If a cause of failure is an invalid input mesh and the algorithm has
provided information on what mesh entities are bad <b>Show bad Mesh</b> 
button appears in the dialog. Clicked, it shows bad mesh entities in
the Viewer in magenta. Sometimes the shown mesh entities are too small
or/and hidden by other mesh elements, to see them it can be helpful to
switch the mesh to Wireframe visualization mode or to switch off
visualization of faces and volumes (if any).

\image html show_bad_mesh.png
<em>Too close nodes causing meshing failure are shown in magenta using <b>Show
    bad Mesh</b> button</em>
<br><br>

\anchor use_existing_anchor
<h2>"Use existing edges" and "Use existing faces" algorithms</h2>

It is possible to create an 1D or a 2D mesh in a python script
(using <em>AddNode, AddEdge</em> and <em>AddFace</em> commands) and
then use such sub-meshes in the construction of a 2D or a 3D mesh. For
this, there exist two algorithms: <b>Use existing edges</b> and <b>Use
  existing faces</b>. Scenario of their usage is following. For
example, you want to use standard algorithms to generate 1D and 3D
meshes and to create 2D mesh by your python code. Then you
<ul>
  <li> create a mesh object, assign an 1D algorithm,</li>
  <li> invoke \b Compute command, which computes an 1D mesh,</li>
  <li> assign <b>Use existing faces</b> and a 3D algorithm,</li>
  <li> run your python code, which creates a 2D mesh,</li>
  <li> invoke \b Compute command, which computes a 3D mesh.</li>
</ul>

Consider trying a sample script demonstrating usage of 
\ref tui_use_existing_faces "Use existing faces" algorithm for
construction of a 2D mesh using Python commands.

\image html use_existing_face_sample_mesh.png
<em> Mesh computed by \ref tui_use_existing_faces "the sample script"
  shown in a Shrink mode.</em>

*/