高级检索

基于中轴的四边形网格生成算法优化

Optimization of Medial Axis-Based Quadrilateral Mesh Generation

  • 摘要: 基于无内点三角网格中轴划分的四边形网格生成算法能够良好地保持模型锐角特征,同时具有流程简洁、计算高效的优点。针对现有算法在处理凹角、倒角和面积剧烈变化区域时,仍存在分块不规则、奇异点难以控制、单元翻转等问题,提出基于中轴的四边形网格生成优化算法。首先分析奇异点分布与无内点三角网格拓扑结构的关系,揭示边界奇异点与顶点邻接Ⅲ类三角形数量的内在联系;然后提出2种预处理方案和基于边交换的中轴分块优化算法,提高分块规则性与边界奇异点度的可控性;最后结合边界优先展平算法和整数规划方法,构建适用于复杂CAD模型的四边形网格生成流程。在多种模型上和主流方法对比实验的结果表明,所提算法在多边界平面、曲面和特征约束CAD模型上均能生成缩放雅可比、最小角等评估指标较高的四边形网格,验证了该算法的有效性。

     

    Abstract: Quadrilateral mesh generation based on medial-axis partitioning of interior-point-free triangular meshes preserves sharp-angle features with a simple and efficient workflow, yet existing methods still suffer from irregular partitions, unstable singularity placement, and element flips in concave, chamfered, or highly varying regions. This work develops an optimized medial-axis-based framework. By examining the link between singularity distribution and the topology of interior-point-free triangular meshes, the analysis reveals a correspondence between boundary singularities and the number of adjacent type-Ⅲ triangles. Two preprocessing strategies and edge-swapping refinement schemes are introduced to improve partition regularity and control boundary-singularity valence. Integrated with boundary-first flattening and integer programming, the approach forms a complete quadrilateral-mesh generation pipeline for complex CAD models. Experiments on planar, curved, and feature-constrained CAD models, together with comparisons against mainstream techniques, report consistently improved mesh quality in metrics such as scaled Jacobians and minimum angles, confirming the effectiveness of the method.

     

/

返回文章
返回