TABLE OF CONTENTS Structured meshes remain quite popular for both two-and three-dimensional simulation applications, primarily due to the ease of developing simulation codes that are based on this topology. It is also generally true that it is simpler to develop mesh tools that operate on structured meshes, as opposed to unstructured topologies. There is not a clear distinction between a structured mesh and an unstructured one ( c.f., Section 1.3), as block-structured and dendritic meshes have properties of both. This chapter seeks to discuss approaches at enhancing structured meshes for simulation purposes, but material will be presented that is also applicable to semi-structured and to particular forms of unstructured meshes. As such, there is some overlap with the following chapter on unstructured mesh enhancement. This chapter will also explore several popular techniques in mesh enhancement and smoothing; many of these ideas also apply to unstructured meshes but are simpler to understand on structured topologies.
The definition of a structured mesh tends to vary somewhat across the various disciplines that employ them. For this discussion, a structured mesh in two dimensions has a straightforward mapping from a parametric ( i, j) computational space to ( x, y) physical space; this mapping (see Fig. 8.1) is generally convenient and usually one to one. A convenient mapping, in this context, is one that is explicitly available or easily constructed. In some cases it might be possible to construct a global logical parametric to physical space mapping for an unstructured mesh, but...
