当前位置: X-MOL 学术Int. J. Solids Struct. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Fully meshfree numerical analysis of nonlocal damage in voxel-based material models from microtomography
International Journal of Solids and Structures ( IF 3.6 ) Pub Date : 2021-03-24 , DOI: 10.1016/j.ijsolstr.2021.03.008
Wei-Jian Li , Qi-Zhi Zhu

Image processing techniques like X-ray microtomography are nowadays extensively applied to reveal micro mechanisms behind macroscopic response of various media and also to reconstruct material microstructure for refined modelling and simulation of diverse properties. For the latter, robust numerical methods making direct use of material and microstructural information at voxels (material grids) and with acceptable computational cost are attracting increasing interests in mechanics and engineering science. In this paper, a fully meshfree iterative computation framework is developed for modelling damage and failure of heterogeneous materials with complex microstructure. The strain field is computed iteratively by mapping operations through a transformation array. The relevant mapping formula is derived from the Lippmann–Schwinger elastic damage problem and operates in the real physical space. On the other hand, the nonlocal damage problem is described by the phase-field method and approximated with a finite difference discretization. Unlike finite element analysis, both the strain and damage fields are approximated iteratively using the same set of material grids, thus unnecessary to solve a system of linear equations. This salient fully meshfree features make numerical computations more efficient than the finite element method and particularly suitable for computing on Graphics Processing Unit (GPU). The computational efficiency and numerical accuracy of the proposed method are demonstrated by various benchmark tests.



中文翻译:

通过显微断层扫描对基于体素的材料模型中的非局部损伤进行完全无网格的数值分析

如今,诸如X射线显微断层照相术之类的图像处理技术被广泛应用,以揭示各种介质的宏观响应背后的微观机制,并重建材料的微观结构,以进行精细建模和各种特性的仿真。对于后者,在体素(材料网格)上直接利用材料和微结构信息的健壮的数值方法以及可接受的计算成本正在吸引越来越多的力学和工程学兴趣。在本文中,开发了一个完全无网格的迭代计算框架,用于建模具有复杂微观结构的异质材料的损坏和破坏。应变场是通过变换数组的映射操作来迭代计算的。相关的映射公式是从Lippmann-Schwinger弹性损伤问题导出的,并且在实际物理空间中运行。另一方面,非局部损伤问题由相场法描述,并通过有限差分离散化近似。与有限元分析不同,应变场和损伤场都使用相同的一组材料网格进行迭代逼近,因此无需求解线性方程组。这种显着的完全无网格特征使数值计算比有限元方法更有效,并且特别适合在图形处理单元(GPU)上进行计算。通过各种基准测试证明了该方法的计算效率和数值精度。非局部损伤问题用相场法描述,并用有限差分离散化近似。与有限元分析不同,应变场和损伤场都使用相同的一组材料网格进行迭代逼近,因此无需求解线性方程组。这种显着的完全无网格特征使数值计算比有限元方法更有效,并且特别适合在图形处理单元(GPU)上进行计算。通过各种基准测试证明了该方法的计算效率和数值精度。非局部损伤问题用相场法描述,并用有限差分离散化近似。与有限元分析不同,应变场和损伤场都使用相同的一组材料网格进行迭代逼近,因此无需求解线性方程组。这种显着的完全无网格特征使数值计算比有限元方法更有效,并且特别适合在图形处理单元(GPU)上进行计算。通过各种基准测试证明了该方法的计算效率和数值精度。因此不必求解线性方程组。这种显着的完全无网格特征使数值计算比有限元方法更有效,并且特别适合在图形处理单元(GPU)上进行计算。通过各种基准测试证明了该方法的计算效率和数值精度。因此无需求解线性方程组。这种显着的完全无网格特征使数值计算比有限元方法更有效,并且特别适合在图形处理单元(GPU)上进行计算。通过各种基准测试证明了该方法的计算效率和数值精度。

更新日期:2021-04-14
down
wechat
bug