一种基于子域分解的表面混合网格生成方法

提出了一种新的基于子域分解的混合网格生成方法。该方法首先用映射法生成结构化背景网格，并确定实体表面上包含的小孔、键槽等小特征在背景网格中的位置，然后删除这些小特征覆盖的背景网格，并在这些区域内生成三角形网格，最后将剩余的背景网格和生成的三角形网格合并，得到整个目标域的网格。该算法综合了映射法效率高、网格质量好、四边形网格计算精度高，以及三角形网格几何适应能力强的优势。数值实验表明，针对复杂的实体表面，新方法能够全自动地生成质量较好的混合网格，生成的网格质量及算法效率均优于传统的推进波前法和铺砖法。

A New Surface Hybrid Mesh Generation Method Based on Subdomain Decomposition

A new hybrid mesh generation method was proposed based on subdomain decomposition. First, the structured background meshes were generated by mapping method and the locations of small features such as hole and keyway were determined. Then, the meshes which covered by small features were eliminated from background meshes, and unstructured triangular meshes were generated in this region. Finally, the meshes of target domain were obtained by combining the background meshes and the unstructured triangular meshes. It is capable of preserving the advantages of mapping's high efficiency and high quality of the meshes, high computational accuracy of quadrangle meshes and the powerful geometric adaptive capacity of triangle meshes. Numerical experiments show that the proposed method may generate high quality hybrid meshes for complex surfaces automatically, the quality of resulting meshes and the algorithm efficiency are superior to traditional methods such as advancing front method and paving.