A NUMERICAL ALGORITHM BASED ON A TREE DATA FOR AN ANISOTROPICALLY ADAPTIVE CARTESIAN MESH

摘要

In an adaptive Cartesian mesh approach, a rectangular mesh is recursively and locally refined and mesh can be automatically generated for complex flow geometry. In this study, a numerical algorithm is developed for an adaptive Cartesian mesh. The tree data is employed to organize the adaptively refined meshes. With this data structure, mesh adaptation becomes very flexible and the algorithm developed for a conventional flow solver can be adapted with less modification. The algorithm can be extended for the anisotropic mesh refinement which is efficient for a boundary layer problem. Parallelization of the developed algorithm is also done in the SPMD paradigm. The domain decomposition technique is used and a tree data structure is split so that computational load should be balanced. Parallel efficiency is examined on a PC cluster.