This command takes two meshes and creates an element-based attribute in mesh1 that contains the number of elements in mesh2 that intersected the respective element in mesh1.

We define intersection as two elements sharing any common point.


intersect_elements / mesh1 / mesh2 / [attrib_name]


[attrib_name] specifies the name of the element based attribute in mesh1 that is created by this command. The default name for this attribute is in_<mesh2>. For example, if the comand syntax was:


the element based attribute that stores the number of intersections would be named in_cmo_well. It is worth noting that GMV does not take kindly to names that are longer than eight characters and will truncate them without even thinking twice, resulting in the name used in our example being changed to in_cmo_w. Therefore, it is good practice to use your own attribute names less than eight characters if possible.

This code has been slightly modified to work with AMR grids produced in X3D. This modification depends on an element based attribute that X3D creates called itetkid. If this attribute is not present, intersect_elements will NOT be able to recognize the AMR grid, and will intersect all elements of the octree. With the itetkid attribute present, only leaves of the octree which intersect will be flagged.

intersect_elements is not designed to work with every element-element combination, but it is pretty thorough. The following table shows what element/element intersetion capabilities are available. An X in the box means that the intersection is supported.

  point line tri quad tet pyr hex point X X X X X X X line X X X X X   X tri X X X X X   X quad X X X X X   X tet X X X X X   X pyr X             hex X X X X X   X ——- ——- ——- ——- ——- ——- ——- ——-

For example, this means that if you have a mesh that has hexes and tets in it, you could intersect it with a mesh that has anything but pyramids in it.

Finally, intersect_elements is based on a k-D-R tree implementation to improve performance in many circumstances. Unfortunately, there is no way to improve performance if the elements being intersected have many candidate elements in their bounding boxes. As such, there are situations where running time may be improved by refining mesh2 such that its elements are of comparable size with those of mesh1.