Generate Quads/Hexes at the Speed of Trias

Quickly constructing a high-quality surface and volume mesh can be a tedious process, especially for complex geometries where sufficient clustering of surface elements along the edges is essential for accurately capturing the flow physics. For such applications, quadrilateral and hexahedral elements have been proven to be useful over triangles or tetrahedra.

To accommodate the needs of such applications, Fidelity Pointwise's mesh generation algorithms enable the rapid generation of quad-dominant unstructured surface meshes. These surface meshes can be used as the basis of a T-Rex (anisotropic tetrahedral extrusion) hybrid volume mesh resulting in boundary layers resolved by layers of unstructured hexahedral cells. The quadrilateral-dominant meshing algorithm in Fidelity Pointwise generates surface meshes consisting of well-aligned quadrilateral cells and does so with the same speed as it takes to generate all triangular meshes. Below are a few examples of unstructured quads/hexes generated using Fidelity Pointwise - 

1. A simple, quad-dominant mesh for a golf club

Figure 1. Quad-dominant mesh for a golf club generated automatically using Fidelity Pointwise's meshing algorithm

2. An ONERA M6 wing was also meshed using the quad-dominant algorithm

Figure 2. Quad-dominant mesh for the ONERA M6 wing generated using Fidelity Pointwise

Figure 3 illustrates fully hybrid meshes with layers of unstructured hexahedral cells in the near-wall region, generated using the surface meshes in Figure 2 (i.e., the surface meshes were used as input to Fidelity Pointwise’s T-Rex algorithm). Relative to prismatic/tetrahedral hybrid meshes, hexahedral/tetrahedral meshes have a lower cell count, improved convergence, and more accurate results. 

Figure 3. A cut-through of the T-Rex mesh for the ONERA M6 wing shows grid cells colored by type (hexahedral cells are blue, prisms are green, pyramids are yellow, and tetrahedra are red).

Sources and Shapes for Control on Local Mesh Sizing

Traditionally, local control over the cell sizes in a tetrahedral mesh away from the boundary layer regions was handled using baffles. A baffle is a piece of topology (for example, a surface mesh) that is embedded inside a volume mesh. The surface mesh of the baffle becomes part of the volume mesh such that small triangle cells on the baffle result in small tetrahedra in the volume mesh. But with the introduction of sources in Pointwise, more flexible control over local mesh sizing is achieved and is explained in the blog Control Cell Size Gradation for Desired CFD Solution Accuracy.

After defining the source's geometry, it is assigned a cell size and an optional cell size variation. The source influences the tetrahedral mesh size but does not constrain it precisely (compared to a baffle where points, edges, and faces become part of the tetrahedral mesh). Furthermore, a source need not be strictly contained within the interior of a block volume. To facilitate the use of sources for local tetrahedral mesh sizing control, Pointwise includes an entity type called a shape. A shape is simply a geometric entity (polyhedron, cylinder, box, and sphere). Figure 4 is an example illustrating the use of a box source behind the DrivAer vehicle to ensure the clustering of cells in the wake region.

Figure 4. The box source (brown) placed behind the DrivAer ensures proper clustering in the wake region.

Other Capabilities to Enhance Meshing Experience

• Expansive Mesh Examination Tools

Support for quadrilateral surface meshes (both structured and unstructured) is available on Fidelity Pointwise, with the realization of the importance of a metric for quad cell warpage. The Warp metric reports the difference in degrees of the two normal vectors that result from diagonalizing the quadrilateral cell because a quadrilateral can be diagonalized in either of two directions, the maximum difference, or worst case is reported. When it comes to understanding how and where your mesh is constrained to the geometry model, the Database Associativity function in Examine provides more insights. 

• Simplified Faceted Geometry

With the varied cell-type support for meshes, it seemed perfectly reasonable to upgrade Pointwise's support for faceted geometry as imported from STL and other files (also known as shells). During import, you have the option to merge coplanar cells into polyhedral or set your desired mix of triangles and quads. This is in addition to the current feature that allows shells to be split along feature lines. The result can be a shell entity that is substantially less dense and, therefore, easier to work with, as shown in Figure 5.

Figure 5. By merging coplanar cells during faceted geometry import you can vastly reduce the size of your shell entities. 

Watch the YouTube video below to learn about unstructured domain algorithms within Fidelity Pointwise - 

The quad-dominant meshing algorithms, the cell-size control via the use of source and shapes, and the additional capabilities such as the mesh examination tools and the simplified faceted geometry within Fidelity Pointwise mesh generation software can enhance user experience and are significant for accurate CFD predictions. 

For more information on quad-dominant meshing/mesh algorithms in Fidelity Pointwise, read Unstructured Quad/Hex Meshes and Mesh Sizing Sources Coming to Pointwise by clicking the button below -