Large scale member experimentation
At the end of the small scale experimentations we saw the most potential in the active bent structures, since here the material and joint together directly instruct the final shape in a clear and intuitive way. Since the small scale experiments were predominantly concerned with understanding and exploring the potential of the three structural systems, the large scale experiments focused on buildability and optimisation. One of the initial questions at the beginning was how the material properties would change and if the active bend structures would still be bendable, since with the increased thickness of the material the initial effort to shape the element would be larger.
Maybe a bit naively we started out just getting some wood, which ended up being pine since we could not argue to ourselves paying a large sum for an exotic type of wood for the initial test. Surprisingly enough, the results we got were nice so we decided to stick with pine as a material. Then it was only beneficial that pine wood has ideal properties regarding its hardness, flexibility, density and elasticity, as well as being a cheap and abundant material.
The design process was iterative, meaning there would be an idea at the beginning. This idea would then be attempted to be built and in the next step laser scanned and and rebuilt as a digital model, leading to a better understanding of how the element would change while setting the foundation for the next experiment.
In practice, this for example looked like this. After having built and scanned the first bent member we were able to visualise how the different versions of the column would look like depending of where along the middle section the elements would be fastened together or which side would be pointing upwards. This then informed the next full scale mock up of the completed column. For its digital reconstruction, a better perception of the space that would make up the top joint was gained. And through a couple of iterations a joint could be developed that perfectly fit into the void.
In the following chapter, I will take you through the step by step process of building the whole structure, starting with the single member and ending with the roof structure.
Column member production
The raw stock material was 20 x 95 mm and roughly 3.3m long rejected pine board. Before building at full-scale, the boards were planed down to a thickness of 10mm to avoid critical points or damaged areas, while also making the piece more flexible. Next, they were all cut to the exact length of 3300 mm and notches were carved out of the sides marking the exact placement of where the rope used for bending and the the jig for construction should be placed. Finally, the surfaces and edges were sanded to avoid injury and allow for the boards to slide easily along each other.
The following drawing illustrates how a member is assembled. After a CNC milled jig is used to drill all the holes in step 1, ropes are placed around the boards at the bottom marks, while a third rope is gradually silt down their length, after they have been forced into a twist by hand, increasing the bend. When all the desired marks have been reached, the boards are fixed in place by putting screws through the predrilled holes.
The design of the jig made of MDF wood allowed for the quick realization of the holes of each board, reducing the time and the level of manual work while increasing its accuracy. Should this be done in an industrial setting, a CNC machine that is able to handle material at this dimension would be used. Same goes for the assembly. Since this project has been done by two people, this means that this whole structure can be built only relying on two people, but on the other hand since some of the steps are quite complex it also requires two people at a minimum without the implementation of special tools for bending and tightening the member.