Mesh Control Entities
Sometimes it is desirable to use certain geometric entities only to control the mesh. For example, you can add a curve inside a domain to control mesh element size there. If you mark this curve as a mesh control entity, it is not included in the geometry used when defining the physics interface and materials. An advantage is that the final mesh need not respect this curve exactly; it is used only to control element size.
Another situation where mesh control entities are useful is when you need precise control of the mesh in certain regions of the geometry. In these regions you typically use a structured mesh with distribution nodes to control the mesh. In other regions of the geometry you can use free (unstructured) mesh.
Suppose that you also want to insert boundary layers. If the boundaries separating the domains with structured and free mesh are ordinary geometry boundaries, the boundary layers have to respect them. This can lead to various problems, including low-quality elements or even meshing failures. If you instead mark such boundaries as mesh control entities, the boundary layer mesh algorithm has more freedom to move mesh nodes and to construct a better mesh.