Adaptive prismatic-tetrahedral grid refinement and redistribution for viscous flows
Adaptive prismatic-tetrahedral grid refinement and redistribution for viscous flows
复制标题
粘性流的自适应棱柱四面体网格细化和重新分布
DOI：
10.2514/3.13131
复制
发表时间：
1996
期刊：
影响因子：
2.5
通讯作者：
Y. Kallinderis
中科院分区：
文献类型：
--
作者：
V. Parthasarathy;Y. Kallinderis
提出了一种适用于三维粘性流动模拟的网格加密与重分布相结合的混合网格自适应算法。流场采用棱柱体单元和四面体单元离散。棱镜区域包括连续的半结构化棱镜层，其包围靠近壁的区域，在该区域中粘性效应占主导地位。棱柱体区域外的栅格是用四面体镶嵌的。提出了一种混合网格自适应方案，实现局部加密和重分布策略，为粘性流计算提供最优网格。混合网格上的网格加密是一种耦合四面体各向同性划分和棱柱方向划分的双重自适应方案。棱镜的方向划分本质上是一个二维网格细化方案，导致所需的计算资源显着减少。网格自适应求解器产生准确的结果相比，一个全球性的细化网格减少计算资源。
A hybrid grid adaptive algorithm that combines grid refinement and redistribution suitable for three-dimensional viscous flow simulations is presented. The flow domain is discretized with both prismatic and tetrahedral elements. The prismatic region comprises successive layers of semistructured prisms that encompass regions close to the wall where the viscous effects are dominant. The grid outside the prismatic region is tessellated with tetrahedra. A hybrid grid adaptation scheme that implements local refinement and redistribution strategies is developed to provide optimum meshes for viscous flow computations. Grid refinement on a hybrid grid is a dual adaptation scheme that couples isotropic division of tetrahedra and directional division of prisms. The directional division of prisms is essentially a two-dimensional grid refinement scheme that results in a significant reduction in required computing resources. The grid adaptive solver yields accurate results as compared with a globally refined grid with reduced computing resources.