The methods described in section MESH do not take into account information about polygonal
and polyhedral
cells contained in a MESH object. Indeed, in the MEDMEM library, the connectivity data for these elements are stored separately . Therefore, the methods that give access to this data are slightly different from those of section MESH.
Also, the polygon and the polyhedra case differ in nature, because in 3D, the list of nodes is not sufficient to described the shape of an element. A descending cell>face>nodes connectivity has to be established to fully describe the elements.
Let us consider the case illustrated in figure fig_polygon_connectivity .
Example for polygon connectivity
For polyhedra, in the nodal connectivity case, one more array is required, because a list of nodes does not suffice to describe a general polyhedron. A general polyhedron is therefore described by a list of faces, each of those being described by a list of nodes.
Let us consider an example with the two tetrahedra represented on figure fig_polyhedron_connectivity , the left one being stored as a MED_TETRA4
element, the right one being stored as a MED_POLYHEDRA
element.
Example for polyhedron connectivity. Node numbers are written with a normal font, while face numbers are written in italic font.
Note that as they are not needed as such, the face numberings are not stored in any array. Only the number of nodes per face is implicitly stored in the polyhedra face connectivity index table.
If there are two MED_POLYHEDRA
elements that share a common face, the list of nodes is repeated twice in the polyhedron connectivity array.
The methods associated to polygons/polyhedra are located in the following classes :
The following example illustrates the creation method for a mesh that contains polygons and/or polyhedra :