Frei Otto Next Generation Architectural Model

This instructable chronicles the process that I followed from fabrication conception to completion of an architectural scale model of an original structure that took major design and engineering ques from Frei Otto's Manhiem Pavilion in Germany.

The concept behind the canopy structure was to take advantage of the structural flexibility of Otto's design of wooden lath structures. Adding on to his design was the design of flexible membranes that would react to the forces applied to the structure itself. These membranes would flex and open depending on the particular stresses, allowing for a more or less porous facade.

From the flattening and laying out of the digital model from a 3d object to 2d flat material, to construction processes and the final product, this model took countless tens of hours and pin pricked fingers.

The final 3d model was designed in Rhino3d and was used to produce renderings, line drawings and physical models.

The structure was generated using Grasshopper and flexed and augmented using the Kangaroo plugin to Grasshopper.

The 3D Model was then stripped apart using another Grasshopper script that took the lengths of each of the segments, converted them to a curve, offset them and capped them with a rounded curve.

The critical component of the dynamic nature of the structure was the precise placement of the joints within the structure.

The joints in this model are made up of 'map tacks' pushed through laser cut holes, then held together with 'earring backs' which hold the pins like the back of an earring holds the back. The smooth metal tack allows the 'wood' (in this case museum board) to pivot around the pin connections smoothly and allows the cells to flex separately.

All of the members are then laid out on a laser cutting template indicating the order of cuts as well as scoring and rasterizing, to be used in a Universal x-660 laser cutter.

Here I chose to cut the holes for the connections before cutting the members themselves out in order to maintain the stable position of the members, so that they were not cut out prematurely and pushed around by the air from the exhaust system.

The same process is used for the 'ETFE' memrane (For the sake of the model a heavy-weight velum was used). In this case the modeling was simpler since only one module was necessary, though this module was repeated over 300 times to complete the final model.

The trick to this step was to double and triple check that the two materials would match up when considering the tolerances of their separate properties and thicknesses. The key is a close and tight fit to ensure rigidity in the z as well as flexibility i the x and y.

Using one sheet of 2-ply museum board and two sheets of Velum all at 18"x32" the members and membranes are cut.

The map tack and earring back combination I used over 300 times to complete this project.

Unfortunately this post is retroactive and the process of building the model was not documented.

The connections were made by sandwiching the velum and two layers of structure together -- the velum was tacked at three sides to the structure while the fourth side was left free to move and not attached to the structure, allowing it to sit down to create an enclosed space, or flex and create a porous envelope.

As it turned out the structure was not strong enough at scale to support itself when it was flexed, so as opposed to the flat design of the Manheim, the model incorporated a truss like system that allowed for more rigidity.

This was an on the fly response to a requirement of more structure that was only discovered upon completion of the original model.

The finished model sat upon a landscape made of laser cut aircraft plywood, and was the center piece of my undergraduate thesis project at UC Berkeley.

For more questions and information please leave a comment or question.

Cheers!