If you’ve ever opened up a three-dimensional modeling application (such as Rhino, 3D Max, Maya, SolidWorks, or Sketchup), chances are you’ve had some experience with ‘Boundary Representations’ (or Breps for short). Whether you knew their technical name or not, almost all Computer Aided Design tools used today are built on the paradigm of Breps, where a “solid” object is that which is enclosed by a set of boundaries or surfaces.
The digital representation of these “solid” objects are often treated as discrete, and homogeneous – where there is a clear distinction between what is “inside” and what is “outside” this boundary envelope. Yet, there are other ways to consider digital material, most notably using a technique called topology optimization.
Topology optimization is a method often used in engineering to produce optimal lightweight structures. We start with a volume of material and redistribute [ie. weaken or strengthen] the material at different points throughout the volume until we achieve an optimal form.
A side effect of this procedure is a suggestion of a structure that is neither hard nor soft but is made out of a ‘virtual material’ of gradually varying stiffness.
Leveraging multi-material 3d printing technology (like that provided by Objet’s Connex multi-material 3D printer), we can directly materialize a new type of ‘fuzzy’ structural system where a soft transparent rubbery material might enclose a series of nested volume of increasing stiffness. These nested harder parts act like an embedded skeleton – but with the transitions between the hard and soft parts blurred.
For the first two experiments we wanted to revisit some known forms and try to reinterpret them through this new prototyping technique [ie. topology optimization coupled with multi-material 3D printing]. Our first example is one of the simplest forms found in all engineering textbooks – the truss. Supporting a long span distributed load between supports at either end, the ghostly outline of a truss is still discernible, while its boundaries are not. In a sense the distinction between bulk and reinforcement is blurred.
The second experiment revisits the leaf as a structure. In nature, a leaf has to cantilever from its stem and carry its own weight while maximizing useful area. Therefore the familiar branching venation pattern of the leaf is the natural answer to this structural problem. Using topology optimization for a slab, the same type of pattern emerges as seen in our example. Again, the reinforcement is not just rigid material encased in soft rubber-like material [something that would be possible with traditional casting techniques] – but instead it is a cascade of nested materials of increasing stiffness and opacity.
This post is also available in: Portuguese (Brazil)