基于弹塑性变形的快速展平算法与网格参数化
A Rapid Unfolding Method Based on Large Elasto-Plastic Deformation and Planar Mesh Parameterization
-
摘要: 针对复杂CAD模型的展平问题, 提出在三角网格上数值模拟汽车覆盖件一步逆成形的优化算法, 实现了对车身零件毛坯料的快速估计.首先对初始网格进行补洞与局部重新网格化等预处理, 得到辅助网格;为了减少变形, 采用基于能量释放的初始解预示算法, 并用基于金属板材弹塑性变形的一步逆算法进行迭代优化, 得到修正的展平网格, 经后处理得到最终参数化结果.将文中算法应用到自由边界网格参数化上, 通过大量数值实验和比较结果表明, 该算法不仅适用于工业上的坯料预估, 而且在网格参数化的面积、扭曲变形方面有明显改善, 同时具有较好的保角效果.Abstract: As a classic case of unfolding the complicated CAD models, this paper presents a rapid optimal algorithm using one-step inverse forming to simulate the process of stamping autobody panels on triangular meshes. Firstly an auxiliary mesh is produced after performing some pre-processing operations, such as hole-filling and local remeshing. An initial guess algorithm based on energy model is then applied to flatten the auxiliary mesh into a planar mesh accounting for shape distortion. The optimization process is conducted to refine the result further based on rapid large elasto-plastic deformation theory, which generates not only a good prediction of the blank shape for manufacturing but also a mesh parameterization with less area distortion and angular distortion. Numerical examples show the effectiveness of our new algorithm.