Introduction

The requirement to create and visualize a complex three-dimensional shape and then discretize the domain by the generation of a computational mesh is a critical operation in computational fluid dynamics. CFD2000 is designed to circumvent the tedium which, historically, has been associated with building high-quality three-dimensional meshes.

CFD2000 provides the tools for constructing and modifying model geometries. The design engineer works with a combination of mouse and keyboard input in a multiple-window environment to create definitions of the physical geometry.  Once the geometric modeling task is completed, CFD2000 separates the definition of the physical geometry from the generation of the computational mesh.  Thus, the edge and surface definitions are retained regardless of the density of the mesh points in the final mesh.

Coordinate Systems

CFD2000 offers a choice among regular (Cartesian), cylindrical-polar, or body-fitted coordinate (BFC) systems—a choice that increases power and flexibility for flow simulations. BFC meshes conform to the local geometric features of the model, making it universally applicable to arbitrary configurations.

Geometric Primitives

CFD2000 imports pre-generated geometries from a library of geometric primitives.  The primitive library can also be expanded by the user.

CAD Data Import

CFD2000 imports CAD data in the form of IGES files.   CAD data can be imported as wireframe, NURBS surfaces, and un-trimmed NURBS surfaces.

Mesh Generation

The numerical solution for any fluid dynamics problem requires that a discrete computational mesh be generated for the physical domain. The characteristics of this mesh have a strong influence on the convergence rate and the accuracy of the solution provided by the flow solver.

Mesh Optimization

The CFD2000 mesh optimization facility improves upon the transfinite interpolation method by carrying the mesh generation process one step further. It uses automatically generated mesh as an initial approximation to a higher quality mesh derived utilizing the technique of elliptical mesh generation. This technique offers advantages over purely algebraic methods:

• good control over the skewness and spacing of the derived grid on surface interiors, while simultaneously allowing complete control over the grid spacing (node distribution) on surface edges

• an ability to produce unique, stable, and smooth grid distributions free of interior maxima or minima (inflection points) in body-fitted coordinates

In addition, elliptical mesh generation works well with irregularly-shaped geometries and can produce meshes that are highly conformal with the edges of individual computational surfaces.