Autonomous Meshing



WCX World Congress Experience
Authors Abstract
Autonomous cars already exist, why should anybody these days spend manual time on mesh preparation? This is a task for a machine, not for a human being. In this session, we will show a one-click way to prepare the mesh for multi-bodies or complex topological objects for 3D printing. The underlying software is already in use for paint shop applications: here it prepares a body in white starting from a CAD geometry fully automatically with 5-8 hours computational time on a desktop machine, while requiring less than 15 minutes of manual work. As an input, tessellated data can be imported from several sources including automatic interfaces allowing to extract the data of multi-bodies from CAD. However, these data are often defective and not manifold. In addition, the describing surface is not represented in an exact way. The only exact information one can rely on at this stage is the position of the vertices of the mesh: they are located directly on the surface. The edges and areas of the triangles may deviate from the original surface. Thus, the vertices are kept and the rest is dismissed or enhanced if necessary. The underlying software uses several techniques to repair, connect and maintain the surface geometry by not only keeping the boundary representation exact but even by enhancing its quality. A volumetric (tetrahedral, hexahedral or hybrid) mesh is then automatically and efficiently constructed from the repaired surface mesh if required. This presentation reveals the techniques enabling the above described functionalities.
Meta TagsDetails
Mokriš, D., Mangold, J., and Schifko, M., "Autonomous Meshing," SAE Technical Paper 2018-01-1386, 2018,
Additional Details
Apr 3, 2018
Product Code
Content Type
Technical Paper