Enhanced strategy for adaptive Cartesian grid generation with arbitrarily complex 3D geometry

This article presents an enhanced strategy for generating Cartesian grids for arbitrary complex three-dimensional (3D) geometry. It addresses the major challenges, including data structures managing surface triangles and Cartesian cells, background Cartesian generation, intersection cell determination, and adaptive mesh refinement. To overcome the problem of inefficiency, we developed a robust method for determining intersecting cells based on the k-dimensional tree data structure combined with the fast 3D-Cartesian box-triangle overlap testing al-gorithm. Furthermore, the minimum distance query algorithm based on the k-dimensional tree was enhanced to improve the query efficiency by reducing the number of backtrackings. The correctness and efficiency of the proposed strategy were evaluated and verified by 3D tests on different configurations. The results show that in terms of mesh generation, the CPUtime/cell was nearly 5.33-5.54x10- 6 s; in terms of the minimum distance query, the query efficiency of the improved backtracking method improved by up to 870.9 times compared with the traversal method. Meshing experiments of several complicated configurations demonstrated that the pro-posed algorithms were accurate, robust, and efficient. At last, an example was examined, and the result dem-onstrates that the mesh generation strategy developed in this study meets the actual flow field calculation requirements.