Development of a physical geometry analysis for manufacturable topology optimized castings.
Final Report Abstract
As part of the project "Development of a physical geometry analysis for process-oriented topology optimization of components for light metal casting", data-based substitute models for evaluating the casting process were successfully developed and integration into existing topology optimization software was achieved. For this purpose, substitute models were developed to describe mold filling and solidification in the die casting process and low-pressure casting process. Ray tracing methods were used for mold filling. In addition, a skeletonization method was developed to be able to make geometry-based statements on solidification and the resulting possible solidification voids in the lowpressure casting process based on the component cross-section. In this context, the medial axis transformation was used to identify defects in cast components in combination with shortest-path algorithms. By further transformation into a graph, the cast components can be automatically evaluated and adjusted with respect to directional solidification. A combined structure-process-optimization framework was developed to integrate the models into the structural optimization process. Finally, spherical neural networks for geometry-independent solidification time prediction were investigated, and their accuracy quantified.
Publications
-
Combining Structural Optimization and Process Assurance in Implicit Modelling for Casting Parts. Materials, 14(13), 3715.
Rosnitschek, Tobias; Erber, Maximilian; Hartmann, Christoph; Volk, Wolfram; Rieg, Frank & Tremmel, Stephan
-
„Optimierung von Metamodell Hyperparametern für die Anpassung an ingenieurtechnische Probleme“, in 19. Gemeinsames Kolloquium Konstruktionstechnik, Sep. 2021.
T. Rosnitschek
-
Geometry-based Assurance of Directional Solidification for Complex Topology-optimized Castings using the Medial Axis Transform. Computer-Aided Design, 152, 103394.
Erber, Maximilian; Rosnitschek, Tobias; Hartmann, Christoph; Alber-Laukant, Bettina; Tremmel, Stephan & Volk, Wolfram
-
Kombination von Topologie-Optimierung und Prozessabsicherung im Leichtmetallguss“, Jahresmagazin Werkstofftechnik : Ingenieurwissenschaften, S. 66–68, 2022.
T. Rosnitschek, M. Erber, W. Volk & S. Tremmel
-
Integrating Geometric Metamodel-Assisted Process Assurance into Topology Optimization of Low-Pressure Die Castings. Proceedings of the Munich Symposium on Lightweight Design 2022, 109-119. Springer International Publishing.
Rosnitschek, Tobias; Erber, Maximilian; Hartmann, Christoph; Alber-Laukant, Bettina; Volk, Wolfram & Tremmel, Stephan
-
Predicting the local solidification time using spherical neural networks. IOP Conference Series: Materials Science and Engineering, 1281(1), 012037.
Erber, Maximilian; Rosnitschek, Tobias; Bauer, Constantin; Ali, Güldali Muhammet; Alber-Laukant, Bettina; Tremmel, Stephan; Volk, Wolfram & Hartmann, Christoph
-
Predicting the solidification time of low pressure die castings using geometric feature-based machine learning metamodels. Procedia CIRP, 118, 1102-1107.
Rosnitschek, Tobias; Erber, Maximilian; Alber-Laukant, Bettina; Hartmann, Christoph; Volk, Wolfram; Rieg, Frank & Tremmel, Stephan
