Reprinting Intelligence: An Exhibit at the World Economics Forum 2018 [1]

[1] Chen, T., Jung-Chew Tse & Shea, K., (2018), ETH, Davos, WEF 2018

3D printed Brain

Table of Content : Abstract | Images | Video


Art installations often refer to three-dimensional works of a sufficient volume and mass that they present non-trivial loading conditions. They may also engage in a level of interactivity with the viewers. Large scale installations are typically either simple in construction, or extremely time intensive to make. From the perspective of mechanical engineering, we detail the computational design and fabrication of a 3D printed brain model that functions as the basis for an augmented reality application. The design requirements are that the brain has a minimum volume of 550x400x300mm and that it must have a non-reflective surface. A stem is designed to provide structural support for the brain. The stem is generated using a set of L-system rules that mimic the pattern of a biological neural network. The resulting frame network is optimized for displacement and stress, with cross sectional size as the optimization variable. The combined geometry of the optimized stem and the brain are segmented in 11 parts to be fabricated separately using Fused Filament Fabrication with Polycarbonate. The parts are assembled using a combination of metallic pins and permanent magnets. The total fabrication time is 22 days, and the final design weighs 5.8 kg. This design process demonstrates a simple way to incorporate aesthetics in the computational design of a functional product. Through segmentation and assembly, it shows the feasibility of printing relatively large scale art installations using a commercial 3D printer with sub-millimeter resolution. With this example, we hope this design process may be adopted by designers and engineers alike.



Generation of the support structure
Generation of the support structure using a set of rewriting rules (a L-system). The grammar interpreter used is Rabbit, a plugin for Grasshopper.
The resulting design is optimized using a deterministic algorithm which minimizes weight while constraining the maximum stress and overall displacement. The optimization variables are the cross section radius of each individual beam member. A volume mesh is generated based on the connectivity and the member radiu using Exoskeleton.
Since the volume of the design far exceeds the maximum printing envelope of the 3D printer, the design is divided into 12 segments. The cuts are made at the groove lines of the brain model. The support structure is split with the brain and grouped together.
The different segments are connected together using a combination of magnets and stainless steel dowels. A circular base supports the design. Center of gravity of the design is outside the circumference of the circular base, and is brought back in by embedding a steel plate at the base.
The fabricated and assembled design is exhibited at the World Economics Forum.