论文标题
HXT的高质量四面体网状
Quality tetrahedral mesh generation with HXT
论文作者
论文摘要
我们在最近的一篇论文（10.1002/nme.5987）中提出了一个快速的3D平行Delaunay内核，用于四面体网状网。但是，由于缺乏必要的网格改进工具，这种内核并不完整。本文以先前的工作为基础，并提出了一个快速的平行网格改进阶段，与替代性开源网格发电机相比，该阶段提供了高质量的四面体网格。我们的网格改进工具包包括拆卸边缘和改进的拉普拉斯平滑，以及一个名为“生长SPR腔”的全新操作员，可以将其视为所有翻转的母亲。本文描述了新的网格改进计划的工作流程以及实施的详细信息。这项研究的结果是一系列称为HXT的开源可扩展软件组件，其总体效率通过两个开源网格发电机的详细比较基准在实际示例中证明了：GMSH和TETGEN。
We proposed, in a recent paper (10.1002/nme.5987), a fast 3D parallel Delaunay kernel for tetrahedral mesh generation. This kernel was however incomplete in the sense that it lacked the necessary mesh improvement tools. The present paper builds on that previous work and proposes a fast parallel mesh improvement stage that delivers high-quality tetrahedral meshes compared to alternative open-source mesh generators. Our mesh improvement toolkit includes edge removal and improved Laplacian smoothing as well as a brand new operator called the Growing SPR Cavity, which can be regarded as the mother of all flips. The paper describes the workflow of the new mesh improvement schedule, as well as the details of the implementation. The result of this research is a series of open-source scalable software components, called HXT, whose overall efficiency is demonstrated on practical examples by means of a detailed comparative benchmark with two open-source mesh generators: Gmsh and TetGen.
