Discretization

Discretization is how your part is turned into the network of tiny points (the mesh) that the solver actually computes with. Peridynamics works on discrete points, so this is the step that turns your smooth geometry into something a computer can simulate.

How the points are generated

The most common option is a random distribution — points are spread through your part. The node type controls how the mesh is created.

Mesh density

The mesh density (sometimes called discretization or resolution) controls how many points are placed in a unit of length. This is usually the single most important setting for accuracy and speed:

• More points — a finer mesh. More accurate results, especially around a crack tip, but a longer run and more disk usage.
• Fewer points — a coarser mesh. Faster and cheaper, but you may miss detail.

A good workflow is to start with a coarse mesh and get a simulation running, then refine it to check that the results don’t change. If results change a lot when you refine the mesh, the mesh was too coarse.

Node sets

For boundary conditions and output, you often need to select a specific group of points — for example, the points along a grip, or the points at a cut. A node set is a named group of points you can create from a file and then refer to by name from other panels. See the Boundary conditions and Output panels.

Generating from a G-code file

Advanced users can build a mesh from a G-code file (for example, one produced by a 3-D printer or CNC machine) instead of a simple box. When overwrite mesh is enabled, PeriHub generates the point distribution from the file’s sampling parameters and dimensions instead of the default geometry.

Tip: don’t set an extremely fine mesh for a first run. It will take much longer to solve and use more disk. Refine only when you need to.