摘要
为高效快速地处理和分析压印成形仿真中大规模网格的渲染和显示,提出不同密度的网格间物理量快速映射算法——最近点法、点邻域法和形函数法。以上三种算法通过在源网格上构建的KD树索引或者使用VTK相关函数快速搜索满足特定条件的节点,以实现大规模数据映射的高效性和准确性。以某纪念币为例,将其分别划分100万和300万四面体网格。通过课题组开发的压印成形仿真软件,使用300万网格进行压印成形仿真分析并获取单元的等效应变和等效应力。为提升软件对物理量数据的快速渲染和显示,采用三种映射算法将300万网格的物理量映射至100万网格上进行后处理数据处理。结果表明:最近点法、点邻域法和形函数法均能成功地映射原始数据的局部特征和全局趋势。为了全面评估映射结果的精确性,引入相对误差分析。使用最近点法映射后x方向位移的最小值和最大值的相对误差分别为0.2%和0.29%。不仅如此,基于VTK最近点搜索算法计算效率为其它方法的十余倍。
In order to efficiently and quickly process and analyze the rendering and display of large-scale mesh data in imprint forming simulation,a fast mapping algorithm of physical quantities between meshes with different densities-nearest point method,point neighborhood method,and shape function method-is proposed.These three algorithms achieve high efficiency and accuracy in large-scale data mappingby quickly searching for nodes that meet specific conditions through the KD tree index built on the source grid or using VTK correlation functions.Taking a certain commemorative coin as an example,it is divided into 1 million and 3million tetrahedral grids,respectively.Through the imprint forming simulation software developed by the research team,3million grids were used to conduct imprint forming simulation analysis and obtain the equivalent strain and equivalent stress of the unit.In order to improve the software's fast rendering and display of physical quantity data,three mapping algorithms are used to map the physical quantities of 3 million grids to 1 million grids for post-processing.The results show that the nearest point method,point neighborhood method,and shape function method can successfully map the local characteristics and global trends of the original data.In order to comprehensively evaluate the accuracy of the mapping results,relative error analysis is introduced.The relative errors of the minimum and maximum x-direction displacement after mapping using the nearest point method are 0.2%and 0.29%respectively.Not only that,the calculation efficiency ofthe nearest point search algorithm based on VTK is more than ten times that of other methods.
作者
董嘉龙
许江平
李阳
余振源
DONG Jia-long;XU Jiang-ping;LI Yang;YU Zhen-yuan(School of Mechanical Engineering,Jiangsu University,Zhenjiang Jiangsu 212013,China)
出处
《计算机仿真》
2025年第11期304-308,325,共6页
Computer Simulation
基金
沈阳中钞造币技术研究有限公司(20220056)
热成形产品成形性能分析及工艺优化研究(20230117)。
关键词
网格映射
插值
有限元数据
最近点搜索
Grid mapping
Interpolation
Finite element data
Nearest neighbor search
