Page 215 - Basic Structured Grid Generation
P. 215

204  Basic Structured Grid Generation

                        boundary edges between nodes will be sides. In other words, boundary integrity may
                        not be preserved, and further steps may be necessary.
                          The advancing front technique, first formulated by George (1971), is an unstructured
                        grid generation method which preserves boundary integrity and has the capacity to
                        create the clustering of high aspect-ratio triangles in boundary-layer regions. In this
                        method, outer and inner (if any) boundary curves of the computational domain, which
                        are commonly defined as piecewise-cubic splines based on a user-specified set of
                        points, are discretized by being divided into straight-line segments which correspond
                        to a chosen node distribution of the boundary of the domain. These sets of straight
                        edges compose the initial ‘fronts’. The fronts then move into the interior of the domain
                        in a ‘marching’ process, in which new points (nodes) and edges are created, old edges
                        are deleted, and triangular elements produced. The vertices of a new triangle consist of
                        the two nodes of a segment of a front and another node either already in the front or
                        newly created. This process continues until there are no edges left in the front, i.e. the
                        front has been annihilated, leaving behind a triangulated domain. Note that the initial
                        choice of nodes on the boundary curves must be strongly dependent on the required
                        grid-cell size, since edges in the initial front will be edges in the final triangulation.


                        8.3.2 Grid control

                        Any grid-generation method should provide for adequate grid control regarding accept-
                        able size and shape of grid cells. The main approach to control of grid-cell size in
                        the AFT (in two dimensions) throughout the computational domain involves first the
                        definition of certain required grid-cell characteristics, and then the generation of a
                        background grid. Control over these characteristics is obtained by the specification of
                        a spatial distribution of appropriate grid parameters over the background grid.
                          The size, shape, and orientation of a triangular grid-cell (see triangle ABC in
                        Fig. 8.19) is roughly described by a set of three independent parameters:

                        • a size parameter δ;
                        • a ‘stretching’ parameter s;
                        • the orientation of the cell φ, which is associated with two mutually orthogonal
                           vectors s, n.

                          To define a grid-cell a user can input four grid generator parameters (δ,s,n x ,n y ),
                        where n x , n y specify the components of n with respect to the axes of the global cartesian

                                                                      C

                                                 y                         d
                                            n
                                                               A            B
                                               f                       sd
                                                           s
                                                    f
                                                            x
                        Fig. 8.19 Descriptive parameters for triangular elements.
   210   211   212   213   214   215   216   217   218   219   220