Postprocessing of Voxel-Based Topologies for Additive Manufacturing Using the Computational Geometry Algorithms Library (CGAL)

Abstract

Abstract : Postprocessing of 3-dimensional (3-D) topologies that are defined as a set of voxels using the Computational Geometry Algorithms Library (CGAL) is discussed. Most topology optimization algorithms return results in a 3-D voxel or 2-dimensional (2-D) pixel representation, which cannot be directly used for additive manufacturing (AM). AM typically uses a surface description of the topology in the stereolithography (STL) format, discretized using triangles. As such, any results from voxel-based topology optimization algorithms must be post-processed before 3-D printing. CGAL is a set of computational geometry algorithms, several of which are suited to the task. The work flow described in this report involves first defining a set of points in space from the voxels, either using their center points or corners. Next, a surface is generated by computing an -hull of the point set. Next, surface simplification can be used to reduce the complexity of the surface without losing any detail. Finally, surface subdivision algorithms, such as Catmull-Clark or Doo-Sabin, can be used to smooth the resulting surface.