Extracting high-fidelity isosurfaces of topology optimized shapes for 3D printing
Quadros, William, Salvatore Elder, Joshua Robbins, Brett Clark
Posters, 25th International Meshing Roundtable, Sandia National Laboratories, September 26-30 2016
25th International Meshing Roundtable
Washington DC, U.S.A.
September 26-30, 2016
William Quadros, Sandia National Labs, US, firstname.lastname@example.org
Salvatore Elder, Cornell University, US, email@example.com
Joshua Robbins, Sandia National Labs, US, firstname.lastname@example.org
Brett Clark, Sandia National Labs, US, email@example.com
Extracting high-fidelity isosurfaces of topology optimized shapes is an important step in additive manufacturing (AM). Unlike traditional design, topology optimization determines material placement for AM based on physics and functional requirements. A topology optimized shape is generally represented using implicit data such as a scalar field on a discrete model (e.g. a density field on a hex mesh). An explicit representation is then generated by constructing an isosurface from the implicit data using a popular algorithm such as marching cubes (MC). The isosurfaces constructed using MC generally contain slivers and abrupt changes in surface normal that cause high-frequency noise. Also, the underlying scalar field itself might have waviness which contributes to low-frequency noise. This high- and low-frequency noise in combination with the stair-step artifact of layer-by-layer AM, results in parts with poor-quality surface finish. Therefore, we propose a method to extract high-fidelity smooth isosurfaces that are suitable for AM using mesh processing operations and shape functions of finite elements. The effectiveness of the proposed method has been visually inspected using zebra stripes and Gaussian curvature plots before 3D printing the parts.
Download Full Paper (PDF)
Contact author(s) or publisher for availability and copyright information on above referenced article