Computing interior support-free structure via hollow-to-fill construction. (February 2018)

• Record Type: Journal Article
• Authors: Yang, Yang; Chai, Shuangming; Fu, Xiao-Ming
• Is Part Of: Computers & graphics. Volume 70(2018)
• Page Start: 148
• Page End: 156
• Publication Date: 2018-02
• Subjects: Interior support-free structure -- Hollow-to-fill -- Shape optimization -- Structural stability -- Fabrication; Computer graphics -- Periodicals; 006.6
• DOI: 10.1016/j.cag.2017.07.005
• Languages: English
• ISSNs: 0097-8493

Abstract

Graphical abstract: Highlights: An approach for generating voxel-based interior support-free structures is proposed. A novel hollow-to-fill strategy is presented to construct support-free inner surfaces. The printed results use less material than the prior methods, and have no seams. Three types of structures are applied to compute support-free inner surfaces. Shape deformation and extra weights are also utilized to optimize the shape. Abstract: Shape optimization often takes into account both inner and outer surfaces to achieve the design goals. The inner surface produced by existing methods is usually not support-free, which results in additional interior support structures during the printing process, thereby disrupting the design goals. In this paper, we present a simple and novel hollow-to-fill algorithm to guarantee the support-free property of inner surfaces for shape optimization. By analyzing the support-free conditions, three types of voxel-based support-free interior structures are proposed to compute inner surfaces. Given a voxelized model and the optimization goal, we first hollow out the model until it becomes a shell, whose thickness is determined by the physical material properties, and then add the support-free structures to optimize the inner surface, and finally refine the inner surface from bottom to top to minimize the optimization objective while maintaining the support-free property. Furthermore, shape deformation and extra weights are also utilized to optimize the shape for design goals. We demonstrate the feasibility and practicability of our method in three modes of balanced object design, where the physical mass properties of rigid objects are satisfied for the purpose of 3D printing.