Mesh generation has evolved rapidly over the last decades and meshing techniques seem to have reached a level of maturity that allows them to calculate complete solutions to complex three-dimensional problems. Typically, unstructured meshes for complex three-dimensional domains of arbitrary shape can be completed on current workstations in reasonable time. Further improvements may still be expected, for instance, regarding the robustness, reliability and optimality of the meshing techniques.

Early mesh generation methods employed meshes consisting of quadrilaterals in two dimensions or hexahedra in three dimensions. Each vertex of such meshes can be readily defined as an array of indices and these types of meshes are commonly referred to as structured meshes. By extension, any mesh having a high degree of ordering (for example, a Cartesian grid) is said to be structured. More recent developments have tried to cope with the complex geometries (for instance, in CAD models involving multiple bounding surfaces) that were difficult to handle (i.e., to mesh) with fully structured meshes. Nowadays unstructured meshes are commonly associated with finite element methods to provide an efficient alternative to structured meshes.

The purpose of this chapter is to provide a comprehensive overview of the current techniques for both structured and unstructured mesh generation and to discuss their intrinsic advantages or weaknesses. These techniques will be further discussed in more detail in the relevant chapters of this book. First, a preliminary classification of existing meshing techniques is proposed. One section is dedicated to surface meshing as surfaces play an important...