中文摘要：

基于混合网格和CGNS（CFD General Notation System）数据结构,建立了一种各向同性加密/稀疏的网格自适应方法。在悬空点的后处理中,让含有悬空点的单元转化为任意多面体,从而简化了自适应单元剖分模版,同时自适应网格单元之间可完全相容,自适应生成的网格能够直接用于可处理任意多面体的流场求解器。将该自适应方法与脱体涡模拟（DES）算法相结合,开展了65°后掠三角翼大迎角流动的数值模拟应用,并与初始网格的模拟结果进行了详细比较。对比表明：采用网格自适应方法适当增加局部网格量,能够以较小的成本迅速提高三角翼背风区的空间分辨率,增强数值模拟对小尺度涡系结构的解析能力,从而弥补了基于混合网格的脱体涡模拟中常用二阶格式计算的空间分辨率相对偏低、不利于湍流多尺度结构精细模拟的不足。

英文摘要：

An adaptive mesh technique with isotropic refining/coarsening approach based on CFD general notation system（CGNS）data structure is presented and implemented for hybrid grids.In order to simplify the possible refinement cases,elements with hanging nodes are changed into polyhedron and meanwhile the compatibility of the grid is maintained,so flow solvers that allow polyhedrons can operate on the adapted meshes without any modifications.Then,detached eddy simulation（DES）method combined with the grid adaptation technique is applied to simulate the flow over a 65°sweep delta wing at high angle of attack.The comparison of the DES results on initial grid and adaptive grid,as well as the experimental data,is carried out.The numerical results demonstrate that with the use of the adaptive technique the spatial resolution in the leeward side of the delta wing can be improved effectively due to the increment of cell number in some local region and the capability of present DES solver to resolve the small scale turbulent flow structure is effectively enhanced,consequently the issue of resolution is alleviated in DES simulation based on hybrid grid with a commonly used second-order scheme.