University of Alabama Birmingham
Geometry- and Image-based Methods for Analyzing Metal Casting Process Features
Abstract
dc:description.abstractThere are no known visual software tools relating metal casting geometry directly to features of the casting process. Solidification simulations do not explicitly disclose defect locations and do not generate rigging geometry. Having geometry-based process feature models would enable cast component designers to improve castability before production. Such models would also allow for a greater degree of workflow automation in casting process optimization. Three papers are written which develop geometry-based process feature models for analysis of metal castings. The first paper explores the relationship between component geometry and solidification order and demonstrates how to rapidly estimate hotspot locations using a segmentation algorithm. The first paper culminates in a practical example validating the method. The second paper builds on the first by publishing and explaining open-source code capable of generating feeders from a CAD model and a mesh density number. The second paper ends with a demonstration of graph-based clustering for feeder aggregation and a discussion of how the software fits into a process simulation and optimization workflow. The third paper builds on the second by expanding the available process feature analyses to include undercuts, parting perimeter information, waterfall, thin-wall, and global geometry metrics. The third paper ends with a discussion of the potential applications for orientation optimization and for machine-learning based castability and cost estimation.
Degree
thesis:*- Level thesis:degree_level
- Dissertation
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Warriner, William
- Contributors dc:contributor
-
- Charles A Monroe
- Robin D Foley
- Amber L Genau
- Leon Jololian
- Mark L Weaver
Subjects
dc:subject × 1Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.library.uab.edu/etd-collection/3274