Skip to main content

MATERIALIZABLE: Intelligent fabrication-oriented Computational Design and Modeling

Periodic Reporting for period 2 - MATERIALIZABLE (MATERIALIZABLE: Intelligent fabrication-oriented Computational Design and Modeling)

Reporting period: 2018-08-01 to 2020-01-31

While access to 3D-printing technology becomes ubiquitous and provides revolutionary possibilities for fabricating complex, functional, multi-material objects with stunning properties, its potential impact is currently significantly limited due to the lack of efficient and intuitive methods for content creation. Existing tools are usually restricted to expert users, have been developed based on the capabilities of traditional manufacturing processes, and do not sufficiently take fabrication constraints into account. Scientifically, we are facing the fundamental challenge that existing simulation techniques and design approaches for predicting the physical properties of materials and objects at the resolution of modern 3D printers are too slow and do not scale with increasing object complexity. The problem is extremely challenging because real world-materials exhibit extraordinary variety and complexity.

To address these challenges, this project suggest a novel computational approach that facilitates intuitive design, accurate and fast simulation techniques, and a functional representation of 3D content. We propose a multi-scale representation of functional goals and hybrid models that describes the physical behavior at a coarse scale and the relationship to the underlying material composition at the resolution of the 3D printer. Our approach is to combine data-driven and physically-based modeling, providing both the required speed and accuracy through smart precomputations and tailored simulation techniques that operate on the data. Subsequently, we propose the fundamental re-thinking of the workflow, leading to solutions that allow synthesizing model instances optimized on-the-fly for a specific output device. The principal applicability will be evaluated for functional goals, such as appearance, deformation, and sensing capabilities.

Linking computation, data, and new digital fabrication technologies may have as profound an impact on the world as the coming of the factory did. New theoretical insight and practical algorithms that extend the traditional content creation pipeline of computer graphics and computer-aided engineering will fundamentally change how functional objects are designed and personalized. Related areas such as robotics, industrial design, architecture, consumer goods creation, and bio printing will benefit because complex designs can be realized more quickly, more accurately, and in a way that was previously impossible. Due to this development, new business opportunities will evolve.
WP1) Modeling, simulation, and reproduction of appearance

We made significant progress towards a computationally efficient model that allows us to simulate the appearance by taking into account lower-scale material phenomena. In contrast to pigment-based colors – where spectral components of incident light are absorbed by the pigment material – structural colorization arises from the interaction of light with micro- and nanostructures. Variations of the refractive index at the length scale of the light’s wavelength cause interference effects that lead to fascinating optical effects such as iridescence. Designing such colorization is challenging, however, since the wave nature of light has to be taken into account when simulating its interplay with nanostructures. Furthermore, realizing such structures requires elaborate methods from nanofabrication. Owning to these complications, linking computational methods and advanced manufacturing is essential for exploring structural colorization, but so far has received little attention from the scientific community. In the work of Auzinger et al. [2], we present a computational design tool for the creation of transparent nanostructures with feature sizes in the range of 100s of nm that realize simple colorization of light transmitted through them. The system is built around a full electromagnetic simulation of light in order to faithfully capture wave- and polarization-dependent effects that approximations to light transport cannot handle. We take fabrication constraints of multiphoton lithography into account to ensure the feasibility of fabricating the designs.

On a larger scale, we also investigated the reproduction of colored textured objects using polyjet 3D printing technology. Color texture reproduction in 3D printing commonly ignores volumetric light transport (cross-talk) between surface points on a 3D print. Such light diffusion leads to significant blur of details and color bleeding, and is particularly severe for highly translucent resin-based print materials. In the work of Elek et al. [7], we counteract heterogeneous scattering to obtain the impression of a crisp albedo texture on top of the 3D print, by optimizing for a fully volumetric material distribution that preserves the target appearance. We evaluated our system using a commercial, five-tone 3D print process, demonstrating that our method preserves high-frequency features well without having to compromise on color gamut. In the follow-up work of Sumin et al. [10], we revisited the assumption that pushing the absorptive materials to the surface results in minimal volumetric cross-talk. We designed a full-fledged optimization on a small domain for this task and confirm this previously reported heuristic. Then, we extended the approach above that is critically limited to color reproduction on planar surfaces, to arbitrary 3D shapes. Our method enables high-fidelity color texture reproduction on 3D prints by effectively compensating for internal light scattering within arbitrarily shaped objects. In addition, we proposed a content-aware gamut mapping that significantly improves color reproduction for the pathological case of thin geometric features. Using a wide range of sample objects, we were able to demonstrate color reproduction whose fidelity is superior to state-of-the-art drivers for color 3D printers.

WP2) Modeling, simulation, and reproduction of elastic behaviour

Many functional objects in our everyday lives consist of elastic, deformable material, and the material properties are often inextricably linked to function. In this reporting period, we made important steps towards modeling systems for designing large, heterogeneous elastic objects with higher-level functional behavior.

FlexMaps [1] is a novel framework we developed for fabricating smooth shapes out of flat, flexible panels with tailored mechanical properties. We start by mapping the 3D surface onto a 2D domain as in traditional UV mapping to design a set of deformable flat panels called FlexMaps. For these panels, we design and obtain specific mechanical properties such that, once they are assembled, the static equilibrium configuration matches the desired 3D shape. Closely connected to this work is self-actuated material and structure design. These types of structures are usually composed of an actuation mechanism and a deformation limiting mechanism that, when coupled together, produce the desired deformed shape. We have developed a computational approach for designing curvy shells that self-actuate from an initially flat state [9].

For reproducing digital objects, we developed a new technique through reusable elastic molds [4]. The molds are generated by casting liquid silicone into custom 3D printed containers called metamolds. Metamolds automatically define the cuts that are needed to extract the cast object from the silicone mold. The shape of metamolds is designed through a novel segmentation technique, which takes into account both geometric and topological constraints involved in the process of mold casting.

Gaining a deeper understanding of the problem, we realized an innovative formulation for how to place cuts in the mold volume to allow for cast extraction [11]. In contrast to the previous work of Alderighi et al. [4], it does not require solving an integer programming problem. Our new formulation allows for elegantly, efficiently, and directly discerning the mold geometry within the volume. The core idea of our solution strategy is based on an analysis of the escape paths of mold volume elements. Intuitively, the escape path of a small mold volume element is the shortest path that this portion of the mold would follow when leaving the cast shape. The behavior of these escape paths can provide information about whether neighboring elements should belong to different mold pieces, and whether they should be separated by additional cuts in the silicone part to ease the extraction. We use geodesic shortest paths to the mold volume boundary as geometric proxies for the escape paths. Our cut configurations are modeled as a set of smooth 3D surfaces, which we call membranes, that can intersect with each other to form a (possibly) non-manifold geometry.

We also recently have developed a general method for shape optimization of models that directly works on CAD model representations and allows to handle complex geometry without the need of remeshing thanks to an underlying XFFEM approach [12].

WP 4) Unified Simulation Framework

In this WP our goal is to develop a framework for multi-objective optimization and efficient design exploration. Working toward this goal, we developed a data-driven technique to instantly predict how fluid flows around various three-dimensional objects [5]. Such simulation is useful for computational fabrication and engineering, but is usually computationally expensive since it requires solving the Navier-Stokes equation for many time steps. We demonstrate the effectiveness of our approach for the interactive design and optimization of a car body. Furthermore, we investigated the interactive design of functional mechanical objects that can be 3D printed [8]. Our computational approach allows novice users to retarget an existing mechanical template to a user-specified input shape. Our system also allows users to efficiently explore various design choices and to synthesize customized mechanical objects that can be fabricated with rapid prototyping technologies. Finally, we made significant progress towards more efficiently reproducing an objects shape based on traditional molding [3]. Our system provides intuitive control for the decomposition of an object into moldable parts. Our technique enables a novel workflow, as it empowers novice users to explore the design space, and it generates fabrication-ready two-piece molds that can be used either for casting or industrial injection molding of free-form objects.
The main scientific result will be solid theoretical insights and novel algorithms that facilitate intelligent fabrication-oriented computational design and modeling. These will enable a new stage for functional digital content creation for 3D printers. The Materializable project will have significant impacts at several levels scientifically and economically. By closing the loop between the real and virtual world, we will be able to scientifically evaluate and improve the models that are currently used in simulation. This will lead to novel, accurate data-driven models that are applicable for analyzing, simulating, and solving various kinds of scientific and engineering problems, thereby having a wide scientific impact on the computer graphics community and beyond. Our theoretical insights and algorithmic foundations for representing functional models, design spaces, and the gamut of output devices will leverage the area of intuitive, functional content creation to a new stage, and we expect that the computer graphics, geometric modeling and processing, and science and engineering community will build future research on this fundament. By disseminating our results in an open-source framework, we will support the scientific and engineering community in this endeavor.