Surface quality improvement and support material reduction in 3D printed shell products based on efficient spectral clustering

For a large-sized free-form surface to be made using additive manufacturing or 3D printing methods, the 3D surface model have to be divided into several parts to be printed in the limited volume of a printing machine. The individual printed piece has to meet requirements of surface quality, using less support material and less printing time. This paper proposes a surface segmentation method to improve surface quality and reduce support material for 3D printed shell products. Surface features related to the surface quality and support material volume are extracted for each triangle on the 3D mesh surface to form a similarity matrix. A graph is built by transferring triangles of the 3D mesh surface into nodes and distances between centers of gravity for two adjacent triangles into edges. Area balance cut (ABcut) is proposed to divide the graph based on the minimum loss function and maximum area in each sub-graph to avoid generating small parts. A Laplacian matrix is structured based on the similarity matrix and ABcut. After clustering the 3D mesh surface into several clusters, common boundaries are optimized to generate smooth boundaries. 3D models for each part are formed and printed by a 3D printer. Case studies validate the performance of the proposed method.