This section describes types of meshing in terms of “patch conforming” and “patch independent” meshing algorithms. In some cases algorithms work at the part level and in others, at the assembly level.
What is patch conforming meshing?
Patch conforming meshing is a meshing technique in which all faces and their boundaries (edges and vertices) [patches] within a very small tolerance are respected for a given part. Mesh based defeaturing is used to overcome difficulties with small features and dirty geometry. Virtual Topology can also lift restrictions on the patches.
As patch conforming mesh methods generally respect all the topology, if a mesh has already been generated, and you change the scope of an object (load, boundary condition, named selection, and so on), the change in the scoping will automatically inherit the associated mesh and the mesh does not have to be re-generated even if Topology Checking is set to Yes.
Note: If significant defeaturing is being done, and objects are scoped to small faces, you need to be careful that the mesh properly captures these boundaries. This is generally best done by putting local mesh sizes on such features.
Mesh Refinement is supported with all of the patch conforming meshers.
Applications
You can implement patch conforming meshing using settings related to any of the following mesh control options, which operate at the part level unless otherwise noted. See the Method Control section for further details.
Solid Bodies:
What is patch independent meshing?
Patch independent meshing is a meshing technique in which faces and their boundaries (edges and vertices) [patches] are not necessarily respected unless there is a load, boundary condition, contact, Named Selection, result, or other object scoped to the faces or edges or vertices (topology). These objects scoped to the geometry instruct the mesher to respect these features; thus, they are called "protected topology." Patch independent mesh methods can do significant defeaturing provided they are able to respect the protected topology, so these approaches are generally useful when gross defeaturing is needed or when a very uniformly-sized mesh is needed.
Notes on protected topology
If surface bodies have differing thicknesses, the edges between the faces will be protected.
Virtual Topology can still be used with patch independent meshing; however, the boundaries of the virtual cells do not have to be protected unless the virtual topology is scoped to something. In other words, the virtual cells replace the underlying geometry but follow the same protection rules.
Topology checks ensure that for each piece of geometry topology, there is properly associated mesh. For example:
Solid bodies should have volume mesh associated to them.
Faces should have face mesh properly associated to them.
Edges should have edge mesh properly associated to them.
Vertices should have nodes associated to them.
The mesh is associated to the geometry even for bodies, faces, edges, and vertices that are not protected, but in such cases there may not be the one-to-one relationship that exists between the mesh and geometry for protected topology. For example, in the figure below, the highlighted edge is protected for the boxes in the bottom row, but not for the boxes in the top row. In the two boxes on the right, the faces with red mesh indicate face mesh that is associated with the triangular face on the left of the geometry. Due to the sharp features of the box, the side faces all have a one-to-one relationship.
After meshing, if a new object is scoped to the geometry, the associations to the mesh may not match (as in the box in the top right corner). In this case, the new object is added to the list of protected topologies, but you may or may not want to re-mesh because the associations may be fine. For example, if you scope a control to one of the side faces of the box, the associated mesh will be fine, but if you select one of the trianglular faces, the associated mesh may be problematic. You can set the Topology Checking control in the Advanced group to control the re-meshing behavior. By default, Topology Checking is set to No, which does not force a re-mesh.

Notes on patch independent meshing
Mesh refinement is not supported with patch independent meshing.
Assembly meshing is a patch independent approach but is implemented differently than part-based patch independent meshing. As described above, “assembly meshing” refers to meshing an entire model as a single mesh process, as compared to part-based or body-based meshing, in which meshing occurs at the part or body level respectively. Assemblies can also be meshed using part-based meshing methods, but in such cases the mesher operates one part at a time, and therefore cannot mesh virtual bodies or evaluate parts that occupy the same space.
Applications
You can implement patch independent meshing using settings related to any of the following mesh control options, which operate at the part level unless otherwise noted. See the Method Control section for further details.
Solid Bodies:
CutCell and Tetrahedrons algorithms (assembly level)
Surface Bodies:
For CutCell and Tetrahedrons algorithms (assembly level):
Faces are protected.
Edges with contacts are protected.
With the exception of edges with contacts, the boundaries of faces (for example, edges without contacts, and vertices) are not protected regardless of the loads, boundary conditions, Named Selections, and/or other objects that may be scoped to them. In addition, these objects are not sent to the solver because the assembly meshing algorithms do not associate the mesh to edges or vertices.
For the MultiZone Quad/Tri mesh method:
Surface bodies with differing material definitions are also protected topology.
Surface bodies with specified variable thickness are not protected. To prevent faces and their boundaries from being meshed over, create an individual Named Selection for each thickness.