ABSTRACT In this work, we propose an efficient ‘r-to-h’ adaptation algorithm for moving boundary problems using only vertex displacements and some local h-adaptation operations. Moving boundary algorithms usually require the regeneration of the mesh to avoid his invalidation in particular when the motion is large and/or several bodies are involved. The proposed approach aims to move the mesh nodes, using the Inverse Distance Weighting interpolation method, while conserving the mesh connectivity until the mesh elements quality deteriorate. At this point, a selective h-adaptation involving local mesh modifications such as edge flipping, local refinement and local coarsening are performed only on the submeshes containing badly shaped elements. The developed approach is valid for multi-body simulations and is able to handle their simultaneous independent displacements. The algorithm implemented is extended to parallel computers to speed-up large 3D moving geometries.

中文翻译：

---

使用并行计算机移动浸入式复杂几何的“R-to-H”网格自适应方法

摘要在这项工作中，我们提出了一种有效的“r-to-h”适应算法，用于仅使用顶点位移和一些局部 h 适应操作的移动边界问题。移动边界算法通常需要重新生成网格以避免其失效，特别是当运动很大和/或涉及多个物体时。所提出的方法旨在使用反距离加权插值方法移动网格节点，同时保持网格连接性，直到网格元素质量下降。此时，仅在包含形状不良元素的子网格上执行涉及局部网格修改（例如边缘翻转、局部细化和局部粗化）的选择性 h 适应。所开发的方法适用于多体模拟，并且能够处理它们同时独立的位移。实现的算法扩展到并行计算机，以加速大型 3D 移动几何图形。