Computational Design of Structured Materials

Advances in manufacturing and processing techniques allow us to create objects with ever increasing complexity. This complexity can be used to create not only aesthetic forms, but also structured materials, whose geometries determine their deformation behavior. Systematic control over the elastic properties is possible with the right structures, but the relationship between structure and elastic properties is nontrivial. This highlights the need for new computational design approaches that can leverage the geometric complexity provided by advanced fabrication techniques, either by providing feedback about the deformation behavior throughout the design process, or by directly translating high-level functional targets into the appropriate material structure. This thesis introduces a novel set of techniques to characterize and create structured materials with various functional targets. We first introduce a data-driven microstructure synthesis approach. After creating a database of small-scale structures that show a clearly defined macroscopic deformation behavior when tiled, we use an interpolation method to create new structures from this database and design objects with spatially varying elastic properties. We then propose a design method to create surfaces with decorative cutouts that integrates aesthetics and stability into a single design process. We combine a discrete element texture approach with a topology optimization method to automatically optimizes the distribution of cutouts, creating a synergy of structure and function. Finally, we develop an approach to characterize the mechanical properties of tessellation-based networks of rods. We create a compact representation of the elastic behavior of such networks and explore the space of a specific type of tessellation, isohedral tilings, to show that we can create a broad range of mechanical behaviors from combinations of simple shapes. We present the resulting mechanical characterization using an intuitive visual representation that lends itself to an efficient exploration of the underlying space of structures