基于AFT-Delaunay的二维解耦并行网格生成算法  被引量:4

A decoupling parallel mesh generation algorithm for two dimensional region based on AFT-Delaunay Technology

在线阅读下载全文

作  者:张宇航[1] 余飞[1] 昌继海[1] 曹杰[1] 关振群[1] 

机构地区:[1]大连理工大学工程力学系工业装备结构分析国家重点实验室,大连116024

出  处:《计算力学学报》2017年第2期191-196,共6页Chinese Journal of Computational Mechanics

基  金:国家自然科学基金(11272074);国家科技重大专项(2011ZX02403)资助项目

摘  要:面向平面任意几何区域网格生成,提出了一种将波前法AFT(Advancing Front Technique)与Delaunay法相结合的解耦并行网格生成算法。算法主要思想是沿着求解几何区域惯性轴,采用扩展的AFT-Delaunay算法生成高质量三角形网格墙,递归地将几何区域动态划分成多个彼此解耦的子区域;采用OpenMP多线程并行技术,将子区域分配给多个CPU并行生成子区域网格;子区域内部的网格生成复用AFT-Delaunay算法,保证了生成网格的质量、效率和一致性要求。本算法优先生成几何边界与交界面网格,有利于提高有限元计算精度;各个子区域的网格生成彼此完全解耦,因此并行网格生成过程无需通信。该方法克服了并行交界面网格质量恶化难题,且具有良好的并行加速比,能够全自动、高效率地并行生成高质量的三角网格。Based on Advancing Front Technique (AFT) combined with a Delaunay method, a new decoupling parallel mesh generation algorithm for two dimensional region is proposed. The main idea of the algorithm is along the inertia axis of geometry domain by using extended AFT-Delaunay algorithm, and the domain is divided dynamically into a plurality of mutually decoupled subdomains. The algorithm uses the OpenMP multi-thread method to allocate the subdomains to multiple CPUs, and subdomain meshes are generated in parallel. The mesh generation within subdomains reuses AFT-Delaunay algorithm, which ensures the quality, speed and consistency of mesh requirements. Since the interface wall mesh is generated firstly, the mesh generation of each subdomain is completely decoupled from each other, so the parallel mesh generation processes run without communication. The proposed method overcomes the mesh quality degradation problem of the parallel interface and has good parallel speedup, which is able to fully automatically generate high-quality triangular mesh with highly efficiency in parallel manner.

关 键 词:并行网格生成 解耦 AFT—Delaunay 墙网格 

分 类 号:O242.21[理学—计算数学]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象