高级检索
曹杰, 赵铭伟, 余飞, 昌继海, 关振群. 高效可靠的边界层网格分块层进生成算法[J]. 计算机辅助设计与图形学学报, 2020, 32(8): 1199-1207. DOI: 10.3724/SP.J.1089.2020.18078
引用本文: 曹杰, 赵铭伟, 余飞, 昌继海, 关振群. 高效可靠的边界层网格分块层进生成算法[J]. 计算机辅助设计与图形学学报, 2020, 32(8): 1199-1207. DOI: 10.3724/SP.J.1089.2020.18078
Cao Jie, Zhao Mingwei, Yu Fei, Chang Jihai, Guan Zhenqun. Efficient and Reliable Advancing Divided-Layer Method for Boundary Layer Mesh[J]. Journal of Computer-Aided Design & Computer Graphics, 2020, 32(8): 1199-1207. DOI: 10.3724/SP.J.1089.2020.18078
Citation: Cao Jie, Zhao Mingwei, Yu Fei, Chang Jihai, Guan Zhenqun. Efficient and Reliable Advancing Divided-Layer Method for Boundary Layer Mesh[J]. Journal of Computer-Aided Design & Computer Graphics, 2020, 32(8): 1199-1207. DOI: 10.3724/SP.J.1089.2020.18078

高效可靠的边界层网格分块层进生成算法

Efficient and Reliable Advancing Divided-Layer Method for Boundary Layer Mesh

  • 摘要: 为了提高复杂3D外形的边界层网格生成速度,提出一种基于离散中轴面的前沿分块层进算法.以闭合的边界三角形网格作为输入,在内部生成只含边界点的约束Delaunay三角化背景网格,聚集所有四面体的外心构成离散的中轴面.根据边界面片离中轴面的距离,将边界面片分为2类:快速推进的开阔区域面片和逐层推进的狭窄区域面片.在Intel Xeon X5650 CPU的单核上对NASA通用研究模型的边界层网格剖分结果表明,该算法的剖分效率是传统层进法或膨胀法的10倍以上,且完成106量级边界面片的分块只用了若干秒.

     

    Abstract: To accelerate boundary layer meshing for 3D complex configurations,an advancing divided-layer method based on a discrete medial surface is presented.For a given surface mesh bounding an arbitrary field,a tetrahedral background mesh only containing boundary nodes is generated by a constrained Delaunay triangulation method,and a discrete medial surface is created with the circumscribed sphere centers of all tetrahedrons.According to the distance from the discrete medial surface,boundary facets are divided into two categories:the open area facets with rapid advance and the narrow area facets with standard layer-by-layer advance.One of boundary layer meshes was generated on the NASA common research model in a single core of an Intel Xeon X5650 CPU.The result shows that the proposed method is over ten times faster than the conventional advancing layer method or the inflating method,and dividing millions of boundary facets only consumes a few seconds.

     

/

返回文章
返回