一种新的准结构网格生成方法

在有限元分析中，高质量的结构网格可以有效地提高有限元分析的精度，但结构网格的几何适应性差，针对复杂边界的二维计算模型，现有的方法很难自动生成高质量的结构网格；而非结构网格几何适应性很好，但存在计算效率低和精度差等问题。提出了一种新的准结构网格生成方法，能够实现复杂区域的网格自动生成并且具有高网格质量。该方法首先对计算区域运用Delaunay三角剖分技术生成粗背景网格；然后利用背景网格，使用优化的Voronoi图生成过渡的蜂巢网格；最后，通过中心圆方法对蜂巢网格单元进行结构网格剖分。分析NACA0012翼型数值模拟结果表明，提出的新准结构网格生成方法能够对边界复杂的模型自动生成高质量的网格，并且通过三种不同拓扑类型网格计算结果相互对比及与实验结果对比，证明准结构网格具有高计算精度。

A novel quasi-structural grid generation method

High quality of a structure grid can effectively improve the accuracy of a finite element analysis,however,limited by the poor adaptability of complex geometry,current methods aiming at automatically generating a high-quality mesh meet a great difficulty for two-dimensional models with a complicated boundary.On the other hand,in spite of the good geometric adaptability of an unstructured grid,low efficiency and low accuracy have become the disadvantages.In this paper,a novel quasi-structural grid generation method is presented.This novel method first implements Delaunay Triangulation to the computational domain to generate a coarse background grid.Then a transitional cellular grid is obtained by using the optimized Voronoi Diagram theory.Finally,the cellular grid is subdivided into a structural grid by center circle method.This new grid generation method is proved to be automatic for complex computational domains and highly accurate for numerical simulations through NACA0012 test cases.