Abstract
Since meshes of poor quality give rise to low accuracy in finite element analysis and kinds of inconveniences in many other applications, mesh smoothing is widely used as an essential technique for the improvement of mesh quality. With respect to this issue, the main contribution of this paper is that a novel mesh smoothing method based on an exterior-angle-split process is proposed. The proposed method contains three main stages: the first stage is independent element geometric transformation performed by exterior-angle-split operations, treating elements unconnected; the second stage is to offset scaling and displacement induced by element transformation; the third stage is to determine the final positions of nodes with a weighted strategy. Theoretical proof describes the regularity of this method and many numerical experiments illustrate its convergence. Not only is this method applicable for triangular mesh, but also can be naturally extended to arbitrary polygonal surface mesh. Quality improvements of demonstrations on triangular and quadrilateral meshes show the effectiveness of this method.
Highlights
Nowadays, mesh casts an irreplaceable role in extensive fields as an effective way of discretization
Since EAS method is based on element geometric transformation, this section compares EAS with other existing state-of-theart element-transformation-based methods which have influence and is widely recognized in the field of mesh smoothing
This paper presents an algorithm for 2D planar and surface mesh smoothing using element geometric transformation
Summary
Mesh casts an irreplaceable role in extensive fields as an effective way of discretization. Meshes with poor quality can lead to precision decline in finite element analysis, distortions in geometry surface modeling and bad effects even failure in graphics rendering. Such situations put forwards higher requirements for mesh quality. This paper presents a mesh smoothing method based on an exterior-angle-split (EAS) process to improve mesh quality. Properties of this algorithm are discussed .
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have