Up to 58 Tets/Hex to untangle Hex meshes

Abstract

The request for high-quality solutions continually grows in a world where more and more tasks are executed through computers. This also counts for fields such as engineering, computer graphics, etc., which use meshes to solve their problems. A mesh is a combination of some elementary elements, for which hexahedral elements are a good choice thanks to their superior numerical features. The solutions reached using these meshes depend on the quality of the elements making up the mesh. The problem is that these individual elements can take on a shape which prevents accurate computations. Such elements are considered to be invalid. To allow users to get accurate results, the shape of these elements must therefore be changed to be considered valid. In this work, we combine the results of two papers to scan a mesh, identify possible invalid elements and then change the shape of these elements to make them valid. With this combination, we end up with a working algorithm. But there is room for improvement, which is why we introduce multiple improvements to speed up the algorithm as well as make it more robust. We then test our algorithm and compare it to another approach. This work, therefore, introduces a new efficient and robust approach to untangle invalid meshes.