Abstract
Historically, computer-aided design (CAD) and computer-aided engineering (CAE) have been developed on fundamentally different representation schemes, on the one hand side, continuous surfaces built on spline basis functions and stitched together in boundary representations discriminating the object’s solid interior from its exterior. On the other hand, finite elements discretizing the object and the respective partial differential equations that need to be solved when simulating physical phenomena. A transition between these representation forms entails many practical issues. Although meshing a CAD model into tetrahedra (short: tets) is considered a solved problem, there is no automatic way back from tets to a meaningful editable CAD model once the tet mesh has undergone topological changes, e.g. in the process of performing topology optimization.
This is not a new observation, the scientific community has recognized this problem years ago and come up with the concept of iso-geometric analysis (IGA) where geometry and analysis is being put on the same (iso) representation scheme: tri-variate splines in most of the cases. Parallel to tensor-product splines, subdivision schemes have been developed and used mostly in film and entertainment industry to model organic shapes because they overcome some of the inherit limitations of tensor-product surfaces, i.e. the ability to represent complex topology in one single control mesh, ultimately defining a watertight solid without the approximation issues implied by trimming spline models.
However, IGA based on subdivision currently is largely limited to subdivision surfaces (shell elements), where it has already shown its benefits. IGA based on subdivision solids is far less explored. This talk will discuss and present subdivision solids as an approach to IGA for objects that do not only consist of a single material but for objects with graded properties as can be manufactured with modern multi-material 3D printing technology.
André Stork studierte Informatik in Darmstadt und promovierte 2000 an der TU Darmstadt. Der Titel seiner Dissertation lautete »Effiziente 3D-Interaktions- und Visualisierungstechniken für benutzerzentrierte Modellierungssysteme«. Im September 2010 verlieh die Technische Universität Darmstadt ihm die Honorarprofessur.
André Stork ist seit 1994 am Fraunhofer-Institut für Graphische Datenverarbeitung beschäftigt; zunächst als wissenschaftlicher Mitarbeiter und seit Anfang 2002 als Leiter der Abteilung für Industrielle Anwendungen (heute: Interaktive Engineering Technologien). Die Abteilung entwickelt innovative Lösungen rund um den virtuellen Produktentwicklungsprozess und adressiert Themen wie Datenkonvertierung, Modellierung (geometrische, mesh-basierte), CAD-CAE-Prozessintegration, Simulation/ Optimierung und interaktive Visualisierung, Kooperationsunterstützung und semantische Informationsverarbeitung. Insbesondere erforscht die Abteilung für die sich rasant verbreitende Technologie des 3D-Drucks bzw. des additiven Fertigens neuartige Konzepte für geometrische Modellierung und Simulation, um das volle Potenzial besser ausschöpfen zu können.
