Block-Structured Adaptive Mesh Refinement on Curvilinear-Orthogonal Grids
Block-Structured Adaptive Mesh Refinement on Curvilinear-Orthogonal Grids
复制标题
DOI：
10.1137/110843940
复制
发表时间：
2012-06
期刊：
影响因子：
--
通讯作者：
U. Ziegler
中科院分区：
文献类型：
--
作者：
U. Ziegler
本文提出了一种适用于磁流体力学方程组的曲线正交网格自适应加密方法。一个块结构的细化方法是使用共享重要的属性与混合中央逆风/约束运输单网格求解器，如角动量守恒的圆柱形和球形网格和磁场的螺线管的保存。自适应网格框架是天体物理流体动力学代码NIRVANA的一部分，该代码由作者不断开发和维护。在MPI库的基础上实现了代码的并行化，并通过空间填充曲线区域分解技术实现了自适应网格模拟中的动态负载平衡。对于各种二维和三维磁流体动力学问题，证明了自适应网格的功能和效率。
An adaptive mesh refinement scheme for the equations of magnetohydrodynamics on curvilinear-orthogonal grids is presented. A block-structured refinement approach is used sharing important properties with the hybrid central-upwind/constrained-transport single grid solver like angular momentum conservation on cylindrical and spherical grids and the preservation of magnetic field solenoidality. The adaptive mesh framework is part of the astrophysical fluid dynamics code NIRVANA which is continuously developed and maintained by the author. The code is parallelized on the basis of the MPI library, and dynamic load balancing in adaptive mesh simulations is achieved by space-filling curve domain decomposition techniques. For various two- and three-dimensional magnetohydrodynamic problems adaptive mesh functionality and efficiency are demonstrated.