空间天气MHD模型的并行自适应和磁场无散计算的初步应用

Parallel AMR Computation on Space Weather

近年来,并行自适应方法受到并行计算学界的普遍关注,特别是在磁流体和流体计算领域。本文将介绍如何在并行自适应工具包PARAMESH的基础上实现磁流体并行自适应计算。我们在工具包的基础上实现了中心迎风差分的CT格式、相应的插值算法和误差控制等功能模块。特别是为了保持磁流体计算所要求的磁场散度为零的限制,我们实现了无散度磁场重构。最后给出的三个磁流体算例表明,我们的计算能自动地捕捉磁流体模型的奇性变化,并可保持散度为零。这一工作将为国内空间天气预测的研究和应用打下良好的基础。

Recently,parallel adaptive mesh refinement(AMR) method is very popular in parallel MHD/HD computation.This paper introduces how to use AMR package-PARAMESH to implement MHD parallel AMR computation.We implement center-constrained transport scheme,corresponding interpolation method,and error estimation controller to accomplish AMR computation.In order to keep the constrain of zero-divergence,we implement divergence-free reconstruction of magnetic field in PARAMESH.Three tests show that AMR can capture the character of MHD model and keep divergence constraint.Our work will contribute to the research and application of space weather forecast in China.