Architectural design: Blowing metal to create chairs and more

The former factory building in the Geistlich district of Schlieren is as cold as ice. Only the robot's welding flame is hot. It silently eats along the sheet metal edge, joining together two identically cut pieces of metal sheet lying on top of one another. It is a process that can be seen in thousands of factories. However, Philipp Dohmen, an architect at ETH Zurich, now reaches for a compressor, connects it to a hole in the sheet metal and inflates the welded structure like an airbed. The only difference is that it needs no plug afterwards. The simple inflation deforms the sheet metal structure, which retains this shape automatically and acquires enormous stability. The latter is considerably greater than shapes manufactured conventionally from laboriously folded metal sheets, since the material can find the optimum shape by itself.

Ordinary commercial metal sheets, extraordinary workmanship

For three years, doctoral student Oskar Zieta and scientist Philipp Dohmen have been researching their technology, known as FIDU (Free Internal Pressure Reshaping), along with Ludger Hovestadt, Professor of Computer Aided Architectural Design (CAAD). One initial product was the "Plopp" chair, which won numerous design prizes such as the Red Dot Design Award. When tested in the ETH Zurich test workshop it withstood 2.5 tons. "That's probably considerably above what is needed in practice and shows how much more the technology is capable of," says Zieta with a grin. However, "Plopp" was only a first step. The aim is to make normal commercial 0.5 -- 2 mm thick sheet metal usable for new applications in architecture via a tool-free deformation method after conventional machining steps such as laser cutting and welding.

Zieta and Dohmen quickly came up against the limitations of the existing infrastructure. The welding robot they were able to use could only machine metal sheets up to a maximum size of 1.50 metres. "That's too small in architecture. One must be able to work at storey height at least," explains Dohmen.

The Chair has now procured a robot that can weld areas of 3 x 6 metres. An external location had to be found due to the shortage of space on the Hönggerberg campus. Finally they were able to lease the former "Boneshed" in the Geistlich district of Schlieren, which was awaiting demolition, for a few years. It took its name from the bones that the Geistlich company processed there to produce glue.

Free forms for load-bearing elements

Zieta and Dohmen now finally have enough space to really get going. "In principle we can now manufacture almost any desired shape. It must simply fit onto one sheet of metal," says Zieta. After countless experiments, the technology has now progressed to the point where free forms are possible for all kinds of load-bearing elements. When the technology is fully mature, architects will no longer be restricted to standard profiles for load-bearing designs. Or in Dohmen's words: "All shapes are standard for us because we will manufacture everything in a seamless digital chain. That will open up entirely new creative possibilities for architects." The metal sheets are cut using a conventional laser, welded together by the robot, and the structure is inflated with an ordinary commercial compressor. This enables components to be prefabricated, taken to the building site stacked on pallets to save space, and then not inflated to attain their final volume until they are in situ.

Basic research project set in motion

The CAAD Chair architects, in collaboration with teams led by Pavel Hora, Professor of Virtual Production, Konrad Wegener, Professor of Production Engineering and Machine Tools, and Peter Uggowitzer, Professor of Materials Science, are currently starting a basic research project to enable the load-bearing ability and deformation behaviour of the structures to be calculated. This is because the material's properties affect the result: the elasticity of the sheet metal, the strain values and even the rolling direction during manufacture influence the shape development.

The structures appear so lightweight that even experts greatly underestimate their load-bearing ability. That is why the architects need universally valid values that make the stability calculable; the plan is for these to emerge from the basic research project as well. For example, a bridge 6 metres long was constructed during a seminar week and was afterwards subjected to a loading test. "We invited structural engineers and asked them what weight the bridge was likely to bear," says Dohmen, " None of them believed it would take more than 200 kilos, maximum 300." The bridge, which weighed 170 kilos, broke when sacks of sand weighing 1800 kilograms were placed on it.

Now, because they finally no longer need to limit themselves to a size of one and a half metres, Dohmen and Zieta can tackle entirely new projects. The first thing they want to do is build rotors for wind turbine generators one size larger than the 1.5 metre diameter prototype that already exists. Designs for crash barriers, structures for passenger compartments on a 1:1 scale and components in general on an architecturally relevant scale -- i.e. one storey high -- are also lined up. The welding robot will have its work cut out.