Particle systems for adaptive, isotropic meshing of CAD models

Area of Science:

• Computational geometry
• Computer-aided design (CAD)
• Finite element analysis (FEA)

Background:

• Generating high-quality surface and volume meshes from complex CAD models is crucial for numerical simulations.
• Existing meshing techniques often struggle with adaptive refinement and handling non-smooth boundaries in CAD data.

Purpose of the Study:

• To develop a novel particle-based approach for generating adaptive triangular surface and tetrahedral volume meshes.
• To enable adaptive meshing that automatically responds to geometric features like curvature and topological boundaries.

Main Methods:

• A hierarchical particle sampling scheme is employed, prioritizing features by dimensionality.
• Particles are distributed by minimizing an energy function in 3D space, with movement constrained to parametric surface spaces.
• The system adapts to curvature and topological separation, creating a smooth sizing field without pre-computed feature sizes.

Main Results:

• The particle-based method successfully generates adaptive triangular surface and tetrahedral volume meshes.
• The approach demonstrates automatic adaptation to geometric complexity, including curvature and boundaries.
• Evaluations show competitive performance compared to existing triangular meshing techniques.

Conclusions:

• The presented particle-based approach offers an effective method for adaptive meshing of CAD models.
• This technique provides a robust solution for handling complex geometries and non-smooth boundaries in meshing applications.