The Re:Vault project is a demonstration of the mutually reinforcing capacities of algorithmic design and robotic fabrication.
While having a latent existence in many historical building practices, most notably of gothic cathedrals, as a formalized process, form-finding techniques have nineteenth century origins in the work of Catalan architect Antoni Gaudi. Gaudi, and the twentieth century exemplar, Frei Otto, carefully crafted intricate materially-performing models - of inverted catenary nets and soap-film surfaces respectively - to study and measure the effect of force upon form. While we would argue that these form-finding processes have always been truly algorithmic in nature the limitation of analogue modes of computation in relation to broader architectural speculation is precisely their fixed relationship to immutable physical laws and material properties. Without the ability to expand, intervene-in, or tune the parameters of form-finding techniques, architects are either forced into subservience to imperfect analogies between these factors and a larger set of extrinsic design intentions, or, must remain content to limit themselves to structural and material investigations. When executed digitally however, form-finding processes are decoupled from the limitations of physics enabling an expansion of the number and nature of negotiable inputs, i.e. constraints.
For this project we have developed custom-written software that allows for the realtime, explicit and open-ended negotiation between structural, fabrication, and assembly influences. The software will eventually incorporate additional constraints that are currently under development. These include passive solar gain, and passive ventilation.
Robotic fabrication is an alliance between generic equipment and custom processes; robots were arguably the first truly open source fabrication tools. The machines themselves, in this case a seven-axis industrial model, are relatively generic, and while continually improving in performance, they are clearly linked to their 50 year old ancestor the Unimate1. The promise of robotic fabrication has always been the ability to perform a multitude of unique tasks from a common programming platform; while specific manufacturers use unique syntax, the offline programming techniques used are consistent. The change in recent years is contextual, a result of the development of computational design processes by innovative architecture practices and educational institutions. The use of algorithms to directly control fabrication tools is a natural progression; it simply requires an understanding of specific fabrication processes and the ability to simulate the kinematics of the machine tool.
Due to the highly contingent nature of Re:Vault's form-finding process conventional generic parts have been replaced with contextualized components. Every piece of the masonry vault and its permanent plywood formwork will be unique and require multi-axis fabrication with most also demanding multiple tool operations. It has therefore been necessary to implement computer-numerically-controlled (CNC) mass-customized fabrication methodologies in direct association with the algorithmic form-generation processes. The instruction code for the pavilion's fabrication, including tool changes, part-nesting and the addition of identification nomenclature is generated directly from the 3D form-found model. The robotic control software used for this purpose has been custom written by Dave Pigram and Wes McGee and has remained in constant evolution in parallel with the robotic fabrication research courses they have taught within the University of Michigan's Graduate Architecture Program. The code thus represents a repository of acquired knowledge, both empirical in the form of material properties and methodological in the form of the specific motion control techniques.
Taubman College's recently acquired seven-axis robotic arm and other computer driven fabrication tools have been fully exploited to make the installation possible. All components of the Re:Vault Pavilion will be fabricated off-site in the school's fabrication laboratory and then transported to the Botanical Gardens for final installation.
The significance of the research is not simply the value of a particular machine, in this case an industrial robot, or a specific fabrication technique. Rather, the research demonstrates a departure away from static object-centric models, with neutral or deterministic relationships to material, toward operative models where a given project's physics - its way of materially entering and occupying the world - is intrinsic to the design process.
Area: 800 sq. ft.
Materials: Robotically cut Berea sandstone and plywood
Maciej Kaczynski, Wes McGee, and Dave Pigram
Consultants: Iain Maxwell, Peter Von Buelow
1 'The 2003 Inductees: Unimate', The Robot Hall of Fame Webpage, The School of Computer Science at Carnegie Mellon University. http://www.robothalloffame.org/unimate.html [Accessed December 5,2010]