/*!
\page pattern_mapping_page Pattern mapping
About patterns
The pattern describes a mesh to generate: positions of nodes within a
geometrical domain and nodal connectivity of elements. As well, a
pattern specifies the so-called key-points, i.e. nodes that will be
located at geometrical vertices. Pattern description is stored in
\.smp file.
The smp file contains 4 sections:
- The first line holds the number of nodes (N).
- The next N lines describe nodes coordinates. Each line holds 2
coordinates of a node.
- A key-points line: indices of nodes to be mapped on geometrical
vertices. An index n refers to a node described on an n-th line of
section 2. The first node index is zero.
- The rest lines describe nodal connectivity of elements, one line
for an element. A line holds indices of nodes forming an element. An
index n refers to a node described on an n-th line of the section
2. The first node index is zero. There must be 3 or 4 indices on a
line: only 2d elements are allowed.
The 2D pattern must contain at least one element and at least one
key-point. All key-points must lay on boundaries.
An example of a simple smp file and a preview of a pattern described
in this file:
\image html image94.gif
Application of pattern mapping
To apply pattern mapping to a geometrical object:
From the \b Modification menu choose the Pattern Mapping item or click
"Pattern mapping" button in the toolbar.
\image html image98.png
"Pattern mapping" button
The following dialog box shall appear:
\image html patternmapping1.png
\image html patternmapping2.png
To apply a pattern to a geometrical object, you should specify:
- a face having the number of vertices equal to the number of
key-points in the pattern; the number of key-points on internal
boundaries of a pattern must also be equal to the number of vertices
on internal boundaries of a face;
- a vertex to which the first key-point should be mapped;
- reverse or not the order of key-points. (The order of vertices of
a face is counterclockwise looking from outside).
Then you either load a .smp pattern file previously created manually
by clicking on the "Load pattern" button, or click on the \b
New button for automatic generation.
\n For an automatic generation you just specify a geometrical face
having a mesh built on it. Mesh nodes lying on face vertices become
key-points. Additionally, you may choose the way of getting nodes
coordinates by projecting nodes on the face instead of using
"positions on face" generated by mesher (if there is any). Faces
having a seam edge can't be used for automatic pattern creation.
When creating a pattern from an existing mesh, there are two possible
cases:
- A sub-mesh on face is selected. A pattern is created from the 2d
elements bound to a face by mesher. Node coordinates are either
"positions on face" computed by mesher, or coordinates got by node
projection on a geometrical surface, according to your choice.
- A mesh where the main shape is a face, is selected. A pattern is
created from all the 2d elements in a mesh. If all mesh elements are
build by mesher, the user can select the way of getting nodes
coordinates, else all nodes are projected on a face surface.
\image html a-patterntype.png
\image html a-patterntype1.png
Mapping algorithm
The mapping algorithm is as follows:
- Key-points are set in the order that they are encountered when
walking along a pattern boundary so that elements are on the left. The
first key-point is preserved.
- Find geometrical vertices corresponding to key-points by vertices
order in a face boundary; here, "Reverse order of key-points" flag is
taken into account.
\image html image95.gif
- Boundary nodes of a pattern are mapped onto edges of a face: a
node located between certain key-points on a pattern boundary is
mapped on a geometrical edge limited by corresponding geometrical
vertices. Node position on an edge reflects its distance from two
key-points.
\image html image96.gif
- Coordinates of a non-boundary node in a parametric space of a face
are defined as following. In a parametric space of a pattern, a node
lays at the intersection of two iso-lines, each of which intersects a
pattern boundary at least at two points. Knowing mapped positions of
boundary nodes, we find where isoline-boundary intersection points are
mapped to, and hence we can find mapped isolines direction and then,
two node positions on two mapped isolines. The eventual mapped
position of a node is found as an average of positions on mapped
isolines.
\image html image97.gif
See Also a sample TUI Script of a
\ref tui_pattern_mapping "Pattern Mapping" operation.
*/