Selective laser melting （SLM） is one of the most popular additive manufacturing （AM） technologies for metal parts. Slicing result, especially for the different dimensional slicing geometry and its topology, plays ...Selective laser melting （SLM） is one of the most popular additive manufacturing （AM） technologies for metal parts. Slicing result, especially for the different dimensional slicing geometry and its topology, plays an important role because of the thermodynamic behavior of metal powders. To get correct geometry and reliable topology, a slicing strategy for SLM is proposed. The unavoidable numerical error caused by sampling and geometric transformation is suppressed firstly, according to shifting the z-coordinate of a vertex with a small value such the shifted vertex is on a slicing plane. The result of vertex-shifting makes it possible to identify different geometric features such as skin surfaces, overhang surfaces, extreme edges and volumetric solid. Second, from geometric primitives a hierarchy of axis-aligned bounding boxes （AABBs） is constructed and used to speed up intersection of slicing planes against sets of triangles. All intersecting segments are given different signs to depict their geometric or topological information. Based the different signs, the different dimensional geometry that is eventually represented by simple and anticlockwise oriented polygons, are identified. Finally, the polygons are classified and nested in a multi-tree data structure set to produce correct topological relations. The result of digital and physical experiments shows the proposed slicing strategy is feasible and robust.展开更多
基金Supported by Key Programs of the Chinese Academy of Sciences(No.KGZD-EW-T0)Research Fund for Scientific and Technological Projects of Chongqing(No.2012gg B40003 and cstc2013yykf C00006)
文摘Selective laser melting （SLM） is one of the most popular additive manufacturing （AM） technologies for metal parts. Slicing result, especially for the different dimensional slicing geometry and its topology, plays an important role because of the thermodynamic behavior of metal powders. To get correct geometry and reliable topology, a slicing strategy for SLM is proposed. The unavoidable numerical error caused by sampling and geometric transformation is suppressed firstly, according to shifting the z-coordinate of a vertex with a small value such the shifted vertex is on a slicing plane. The result of vertex-shifting makes it possible to identify different geometric features such as skin surfaces, overhang surfaces, extreme edges and volumetric solid. Second, from geometric primitives a hierarchy of axis-aligned bounding boxes （AABBs） is constructed and used to speed up intersection of slicing planes against sets of triangles. All intersecting segments are given different signs to depict their geometric or topological information. Based the different signs, the different dimensional geometry that is eventually represented by simple and anticlockwise oriented polygons, are identified. Finally, the polygons are classified and nested in a multi-tree data structure set to produce correct topological relations. The result of digital and physical experiments shows the proposed slicing strategy is feasible and robust.
