Slice coherence in a query-based architecture for 3D heterogeneous printing
Graphical abstract
Introduction
Object properties and performance characteristics depend, to a large extent, on the structure of the object’s interior. For instance, the strength of polycrystalline material cannot be determined without information about the shape and orientation of the interior cells, and their material composition. The outer boundary of the object normally determines the geometry of the object and reflects intended functionality, but the functionality must be supported by appropriate interior structure. In particular, the structure of the interior determines properties such as density, elasticity, stiffness, fatigue, current flow, sound absorption, thermal conductivity, and other salient properties of the material.
In order to analyze and study these properties in detail, many researchers model the interior structures of the objects as an aggregation of cells. A cell structure based on Voronoi tessellation is one of the most popular approaches in these studies, since it can generate realistic homogeneous and heterogeneous structures in both 2D and 3D. In fact, Voronoi tessellations can be considered to subsume many types of cell decomposition where the Voronoi sites are arranged in a special way. For instance, when the sites lie on the vertices of a regular grid we obtain a regular subdivision into blocks. In the following we assume a general position site arrangement unless otherwise noted.
We consider representations in which the volume is partitioned into cells where each cell has an interior with specific geometric and/or material specifications. For example, a particular cell might be stipulated to have
- •
a particular strength in a given direction,
- •
or a particular regular geometric structure of a certain characteristic,
- •
or be a specific, homogeneous material, and so on.
We investigate a manufacturing approach in which cells are separated by a membrane structure, and the interior is fabricated using filament deposition equipment or other fabrication techniques that proceed in the usual layer-by-layer methodology. We restrict to manufacturing the membrane separating the cells of the object partition. Specifically, we assume a cell geometry defined by a 3D Voronoi partition of the interior. The Voronoi partition is defined by a set of interior points, the sites of the partition. The sites may be specified explicitly as part of the design. Each such point has attributes that are a description of the cell interior. The attributes are sufficient to unambiguously specify cell interior. Examples include stipulated material, homogeneously filling the cell; variable density material where the density at a point x in the cell might be a function of the distance of the point from the cells site; empty cells; and so on. Cells are therefore convex polyhedra, except for cells that share part of the exterior boundary of the object. The representation is implicit in the sense that the cell specification is given by the sites along with the outer geometry of the object.
We will describe how to manufacture the cell membrane throughout the interior. Our algorithm is incremental, exploiting the fact that many successive slices are topologically the same, and that slices that differ topologically often differ only locally. Those differences are characterized by only a few cases. More than that, the algorithm seeks to deposit each membrane layer such that, at no time, the head requires repositioning without depositing material, unless the external boundary requires it.
The remainder of the paper is structured as follows. After a review of the literature is presented in the next section, the algorithm for our proposed manufacturing paradigm is given in Section 3. The paradigm is evaluated with various experiments and the results are summarized in Section 4. The paper concludes with some remarks in Section 5.
Section snippets
Background
Voronoi based modeling is used in different fields of science and engineering for many applications. Material scientists employ Voronoi tessellations to model and simulate polycrystalline materials under different conditions to analyze their physical properties. In these simulations, each Voronoi cell corresponds to a crystalline cell in the aggregate and the cells form the basis of finite element models used to determine elastic and plastic behaviors of the so modeled materials [1].
Besides
Print setup
In our experiments, we have decoupled the link between the CAD file and the printer, bypassing the proprietary software that creates the STL file. We generate the printer (machine) code directly and drive the printer from a Raspberry micro-computer that requests the G-code for the STL slices one-by-one, when the printer is ready for the next slice. This setup, although for this paper unnecessary, was chosen in anticipation of future experiments in which the dialog between printer and CAD system
Experimental evaluation of the manufacturing paradigm
We evaluate the performance of the proposed manufacturing paradigm for 3D printers using a MakerBot Replicator 2X printer. The experimental set-up is illustrated in Fig. 7. The printer has two nozzle heads and is directly connected to a single-board computer (Raspberry Pi) through USB. The single-board computer operates as a server sending slice files to the printer upon request, the printer being the client for this interaction. In turn, the raspberry requests the slice files from the CAD
Discussion
The traditional approach of treating each slice as a separate structure ignores the fact that, in many cases, there is a high degree of similarity between consecutive slices, geometrically as well as topologically. This is trivially the case when the interior of the object printed is filled with a fixed pattern that has little functional significance. A structure derived from a 3D Voronoi diagram, on the other hand, has functional significance, as evidenced by the extensive literature on the
Acknowledgment
The project has been partially supported by the National Science Foundation, CMMI-1361783.
References (27)
- et al.
Two scale analysis of heterogeneous elastic–plastic materials with asymptotic homogenization and voronoi cell finite element model
Comput Methods Appl Mech Engrg
(1996) - et al.
Quantitative morphology analysis of polymers foamed with supercritical carbon dioxide using Voronoi diagrams
Comput Mater Sci
(2007) - et al.
Modelling fruit (micro)structures, why and how?
Trends Food Sci Technol
(2008) - et al.
Three-dimensional periodic voronoi grain models and micromechanical fe-simulations of a two-phase steel
Comput Mater Sci
(2002) Thermo-elastic analysis of functionally graded materials using voronoi elements
Mater Sci Eng A
(2001)- et al.
Heterogeneous object modeling: A review
Comput-Aided Des
(2007) - et al.
Characterisation of models of multicomponent amorphous metals: the radical alternative to the voronoi polyhedron
J Non-Cryst Solids
(1982) - et al.
Quadruple nodes and grain boundary connectivity in three dimensions
Acta Mater
(2014) - et al.
Functionally heterogeneous porous scaffold design for tissue engineering
Comput-Aided Des
(2013) - et al.
3D voronoi skeletons and their usage for the characterization and recognition of 3D organ shape
Comput Vis Image Underst
(1997)
Geometric interoperability via queries
Comput-Aided Des
An image analysis method to obtain the effective thermal conductivity of metallic foams via a redefined concept of shape factor
Appl Therm Eng
3D virtual pome fruit tissue generation based on cell growth modeling
Food Bioprocess Technol
Cited by (30)
A heterogeneous infill technique for fused deposition modeling
2020, Procedia ManufacturingA continuous path planning approach on Voronoi diagrams for robotics and manufacturing applications
2019, Procedia ManufacturingAnisotropic porous structure modeling for 3D printed objects
2018, Computers and Graphics (Pergamon)Citation Excerpt :One group of methods employ stochastic functions to model the structure close to the existing porous media, using simulated annealing [21], a pressure correction algorithm [22], Monte Carlo methods [23], CSG operations [24]. Another group of methods follow a Voronoi-based stochastic modeling manner [25–28]. Voronoi diagrams are used for tessellating the given shape in a stochastic but space-pleasing manner, and then the pores are modelled via inscribed implicit curves or parametric B-splines.
Tuning the Center of Gravity of 3D Printed Artifacts
2018, Procedia ManufacturingThe Role of Additive Manufacturing in the Era of Industry 4.0
2017, Procedia ManufacturingImplicit slicing for functionally tailored additive manufacturing
2016, CAD Computer Aided DesignCitation Excerpt :One such slicer utilizes General-Purpose Graphics Processing Unit (GPGPU) computing [44]. Additionally, slicers intended to produce objects with spatially-varying stiffness properties have also been developed [45,46]. In the present work a standard STL file is used for input, due to its widespread use in the additive manufacturing industry.