Fabrication Materials Proposal

We looked at several materials used in combination to create a wall system. The first layer is mainly used for creating the geometry, using the same method of pulling and tensioning. The material is then hardened to shape and would act as reinforcement in the wall. Layers of materials could be added to achieve different performative qualities within the wall. 

Shotcrete reinforced with Barchip fibres would be used on the surface of the added materials. Barchip reinforced concrete has good compressive and tensile qualities. We plan on testing several materials for the next steps to get closer to a combination of materials that we can use to test at a 1:1 prototype.

DigiHYBIOS 1.0 & AnaHYBIOS 3.0 / New Approach: From Digital FormFinding to Analogue Methods

With the latest HYBIOS, we adopted a new approach. This time the experiment was based on a digital FormFinding algorithm using Daniel Piker's Kangaroo Plugin for Grasshopper. The new experiment also resolved some of the issues of the previous experiment. 

A cellular grid was used to generate a frame for pathways to have a more even floor to walk on. This time, the modeling was done to scale. Points were restricted at the pathway edges to create ramps on the raised platform. The mesh was relaxed within the grid cells to create blob-like forms by restricting the outlines of the grid. The roof membrane was stretched towards the center of the cells. The idea is to create a circulation path within the space, where depressions within the grid cells could be used for seating etc. Although more control was gained over the digital FormFinding process, there were a lot of restrictions with using the algorithm. Material behaviour and mesh qualities could not be embedded within the logic of the algorithm. Deforming the mesh was not done in real time either, which was a major restriction. We would like to create a script eventually that solves the problems of the current tools we are using and calculates the forces as output.

We deployed the same logic with the analogue prototype. The results were approximately similar, although we have more control over the shaping process, which enabled us to generate more interesting archi-tectonics. The idea of the developed formwork is that the floor grid could be reused to cast several other HYBIOS. The model was constructed to a scale of 1:50 for a 20m x 20m space. Essentially the point of this experiment is to control the reusable formwork to form the structures.


New Approach: From Digital FormFinding to Analogue Methods

We looked at the paper entitled 'Linking Hanging Chain Models to Fabrication' by Axel Kilian as a source of information and inspiration to carry out our next experiment. The next experiment mainly had to focus on utilizing digital form finding methods to add to our methods for producing the next prototype. 

Axel Kilian, Ph.D. Candidate in Computation, Department of Architecture, School of Architecture and Planning, Massachusetts Institute of Technology, Cambridge, MA.

Fabrication output is an integral part of the iterative process and not a post design process.
The relationship between form finding model and the translation into volumetric form was explored in a series of small models.

His Motivation :

Equilibrium solutions can be scaled if the proportional distribution of mass is kept and the geometry of the lines of forces is scaled proportionally. This holds true even though mass does not scale proportionally to geometric dimensions.

A major disadvantage is that a physical model is hard to measure accurately and
in reasonable time, as measuring requires physically accessing the model. The
measurement of forces within the strings of the model is even more difficult, as it
requires the installation of strain gauges, which is time consuming and can potentially
disturb the model. In addition, the measurements are not part of the design process. The
design is frozen to allow for iterating through the load measurements throughout the
model in one given state

The digital version, in contrast, allows simultaneous measurement and creating/editing of
geometry. These measurements can directly drive other dimensions in the model. In the
digital model, editing and creating the string weight is less limited by the availability and
preparation of the physical material, which in the case of a very complex model can
substantially slow down the process. Furthermore, the use of generative techniques
allows for the rapid placement of complex string constructs and observing their behavior
before investing time into an elaborate physical model

The self-weight of the load-bearing member contributes only negligible
amounts to the structure locally and therefore does not substantially affect the hanging
curve form. If there is no load present other than the weight of the structure itself, the
self-weight becomes the dominant form giving factor. The cross section has to provide
enough area for the forces traveling through it. A further optimization of the structure, for
example with the aim of achieving uniform compressional loading throughout the same
material, which would be possible by varying thickness, was not undertaken by Gaudi.

Islers Conclusion : page 8
Isler identifies instabilities in his shells as follows: First, at the
supports second, due to general buckling; third, due to local buckling of the free edge (for
which the counter curvature is so important)

Designing in dynamics vs. analytical approach. Design by discovery : Page 10

Structural and dimensional evaluation of form is not an afterthought but an essential of
the design process. This is where the learning by discovery enabled by interactive tools comes into play. In interaction with a live, force-geometry linked structure, a designer can directly observe the range of structural responses while exploring possible forms. This encourages an explorative approach to design and supports unconventional solutions that integrate and respond to the designer’s intent.


1. form finding,
2. topology finding,
3. load path finding,
4. material distribution
5. testing for structural redundancy
6. optimization.

Form finding techniques in an interactive digital modeling environment can support the
design process by giving continuous feedback to the designer, allowing the designer to
integrate structural principles into the creation of form rather than to structurally optimize
the finished form at the end of the design process.

Analysis of Previous Experiments / Prototypes

In order for us to move forward with the project in a more effective manner, we started with analyzing what we have done before. The previous prototypes (Hybios 2.0 & 2.1) were analyzed mainly for the methods used and the resultant geometries obtained by these methods. The results and failures of each prototype were documented to give insight for the problems that need to be solved in the next prototype.

The main intention of documenting the choreography of the random stretching procedures is to obtain a draft for a set of instructions, or a 'pseudo code' for the construction team to carry on the construction.

A quick experiment to obtain a clearer mapping of the curvatures of the generated topologies was carried out. A grid was projected onto the surface of the prototypes. The intention being to give purpose for the these curvatures rather than just being manipulators of the internal space. Possible environmental responses such as sheltering canopies at entrances and channeling rainwater were discussed.