7.13. Rezoning Restrictions

After you initiate the rezoning process (REZONE), and until you map the solved variables from the old mesh to the new mesh (MAPSOLVE) at the conclusion of the rezoning process, you will lose your rezoning work if you exit the solution processor (/SOLU) prematurely. This condition applies even if you save (SAVE) before exiting and then resume (RESUME) after returning to the solution processor.

Issuing a REMESH,READ command (to import a generic new mesh) or a REMESH,SPLIT command (to split an existing mesh) causes the mesh to lose associativity with its corresponding geometry. It is therefore not possible to issue commands that perform geometry-based operations. (You can issue commands that perform mesh-based operations, however.)

When using a generic new mesh created by a third-party application (REMESH,READ):

  • Mesh regions cannot intersect.

  • If multiple regions are read in (REMESH,READ) via a single .cdb file, the nodes of these regions must be encapsulated in a single NBLOCK environment, and the elements must be encapsulated in a single EBLOCK environment.

For a 2D analysis, you can issue the following remeshing commands in the same session on multiple regions:

AREMESH,AMESH (to create a program-generated new mesh)
REMESH,READ (to import a generic new mesh)
REMESH,SPLIT (to split an existing mesh)

For a 3D analysis, remeshing is possible only by importing a new mesh created by another application (REMESH,READ) or by splitting an existing mesh (REMESH,SPLIT).

Components are not passed through the rezoning operation.

Rezoning does not support:

  • POST1 load-case combination commands (such as LCASE, LCDEF, LCOPER, and LCSEL)

  • Initial state loading (INISTATE)

  • Dropped midside nodes for higher-order elements (EMID)

  • Coupling (CP) and constraint equations (CE)

  • Cyclic symmetry boundary conditions

  • Distributed-memory parallel processing