Tetrahedral mesh segmentation based on quality criteria

Abstract

Simulations based on the Finite Element Method are widely applied in different contexts. The convergence and reliability of results depends directly on the quality of the tetrahedrons that compose FEM meshes. In this context, this work aimed to obtain tetrahedral meshes of real structures from tomographic images using open source software and split those tetrahedrons that do not contribute for convergence of such simulations through the aspect ratio criteria. As a result, tetrahedral meshes from real structures were obtained together with indications of the regions that could impair FEM simulations. Moreover, this approach makes it possible to study specific methods for local refinements for improving numerical simulations. These findings are of great interest to users generally, and particularly to researchers working on geometric and topological methods for shape and solid modeling, with extensive applications in the fields of Computational Fluid Dynamics, Heat Transfer and others.