tet_mesh_refine, a C++ code which refines every tetrahedron in a tet mesh, replacing each tetrahedron by 8 subtetrahedrons.
In particular, the code reads two files, one listing the coordinates of nodes, the other the indices of nodes forming tetrahedrons. From this information, it determines a refinement of the mesh, and writes out new node and tetrahedron files corresponding to this refinement.
The refinement of a single tetrahedron is easy. Every pair of nodes generates a new node whose location is the average of the locations of the original two nodes. The problem is that many tetrahedrons may share the same edge. The new node needs to be generated exactly once, and assigned a unique index, which all the tetrahedrons will share. The tricky part of this calculation is thus figuring out whether it is time to generate a new node, or whether the node has already been generated, in which case the appropriate index must be retrieved.
The current version of the code can only refine a tet mesh of order 4. There's no real reason why it could not also handle meshes of order 10, except that the programming is a bit more complex. For the moment, a mesh of order 10 could be handled by using the sequence of programs:
TET_MESH_Q2L --> TET_MESH_REFINE --> TET_MESH_L2Q
tet_mesh_refine prefixwhere prefix is the common file prefix:
The element definition file will list node indices. In C++, it may be more natural to use 0-based indices. This program will accept an element definition file that is 0-based or 1-based, and will convert a 1-based input file so that it becomes 0-based internal to the program. The detection of 1-based data is determined by the absence of the use of a 0 index, and the use of an index equal to the number of nodes. This is an implicit and fallible, but reasonable, way to handle this problem.
The input and output files use the simple TABLE format; comment lines begin with a "#" character. Otherwise, each line of the file contains one set of information, either the coordinates of a node (for a node file), or the indices of nodes that make up a tetrahedron, (for a tetrahedron file).
The input file nodes.txt contains the coordinate information for the tet mesh. Each data line contains the X and Y coordinates of a single node.
The input file tetras.txt contains the tetrahedron information for the tet mesh. Each data line contains the indices of the nodes that form a tetrahedron.
The output files have the same format.
The computer code and data files described and made available on this web page are distributed under the MIT license
tet_mesh_refine is available in a C++ version and a FORTRAN90 version and a MATLAB version.
TET_MESH, a C++ code which includes a variety of routines for working with tetrahedral meshes.
TET_MESH_BOUNDARY, a C++ code which returns the nodes and faces of the boundary of a tetrahedral mesh, which themselves form a 3D triangular mesh or "TRI_SURFACE".
TET_MESH_DISPLAY_OPENGL, a C++ code which reads a tet mesh and displays the nodes and edges using OpenGL.
TET_MESH_L2Q, a C++ code which takes a 4-node tet mesh and makes a 10-node tet mesh by adding nodes.
TET_MESH_ORDER4, a directory which contains a description and examples of a tet mesh using order 4 elements.
TET_MESH_ORDER10, a directory which contains a description and examples of a tet mesh using order 10 elements.
TET_MESH_Q2L, a C++ code which takes a 10-node tet mesh and makes a 4-node tet mesh.
TET_MESH_QUALITY, a C++ code which takes a 4-node tet mesh and computes the "quality" of the mesh.
TET_MESH_RCM, a C++ code which takes a tet mesh and relabels the nodes to reduce the bandwidth of the corresponding adjacency matrix.
TET_MESH_TET_NEIGHBORS, a C++ code which computes the tetrahedral adjacency information.
TET_MESH_VOLUMES, a C++ code which computes the volume of each tetrahedron in a tet mesh;