Back to boxes

The final step of the project investigates how a curve roof structure could be constructed in a way that mimics the geometry of the ruled surface, but aligns to a horizontal structural system.

Ruled Roofs & Pathways

In search of a more sophisticated system that reflects my early design agenda, I have experimented with ways of using a ruled roof surface in conjunction with a circulation pathway across the camp. Initial investigations show there are a variety ways this could implemented, creating either a single gesture or a more broken roofscape.

A more modular roof feels conceptually stronger when considering the infinite possible arrangements of the camp that can be generated from the script. After developing a system to implement this and dictate the way the roof surfaces are generated, below shows four examples derived from different locations, configurations, and possible pathways .

Architectural ‘fun’ can now be had with the exploring the way these structures are designed. What happens when the two systems, pathway and roof intersect? How is this resolved in construction?

Ruled Roofs

I have begun exploring the potential application of ruled surfaces, the most obvious starting point being as roof or roofscape across the camp. The generated configuration that was reached by script is used as an example scenario to test ideas.

It is easy to see how an architectural conclusion can be reached, and could be resolved quickly from this point. However, I feel such a proposal does not yet fulfill the complexities of the project as intended, and I wish to explore a more sophisticated roof/circulation system using some of these ideas.

Ruled surfaces: curves from lines.

Given the orthogonal nature of the project so far, ruled surfaces seemed a natural place to begin an investigation into curved forms.

In 3D model and grasshopper, I investigated how varying surfaces can be created following different ‘stem’ lines on the faces of a cube.

In attempt to classify the resulting surfaces I sought a graphical way of analysing their curvature, that could be represented and compared between each. As shown in the diagrams below, the mid-point curve is generated in two directions across the surface. The resulting curves are laid flat in mirror of each other, and area of each calculated as a percentage of the overall polygon from endpoints. Because the ‘curvy-ness’ of the surface increases as the mid-point curve deviates from a straight line, the large the percentage area is deemed to represent a more curved surface (…maybe…).

String models of 6 example surfaces.

Dropping Boxes: accessibility analysis

Taking a step away from an entirely scripted system, I have been considering accessibility and circulation within and across the ‘dropped’ boxes. This has involved introducing rules to classify ‘accessible’ vs ‘inaccessible’ boxes, based on max. permissible step levels and min. number of touching boxes. The inaccessible boxes are then removed from the grid creating potential breaks or open spaces within the camp. It is noted that structures dropped with a high snap threshold value do not loose any boxes, as due to the alignment none become regarded as ‘inaccessible’.

The second part of this analysis looks at how the locally accessible blocks could exist as distinct units, removing the walls between their constituent boxes. However, as accessibility between them is not possible due to the max. permissible step value, a second tier of circulation must exist. It is thought that this could exist as part of the rooftop structure, with the introduction of steps/ladders/stairs between the units.

It is at this point that the project may move into a more typically ‘architectural’ design scheme, and away from the initially scripted geometry.

Dropping Boxes: resolving form with a snap resolution.

In considering an eventual architecture to the system, it became apparent that a structure based on precise terrain values would be impractical to design, draw and build. At some point, regardless of material and structure, there needs to be a ‘threshold’ of height variables that hypothetical boxes can rest at. To analyse this I introduced a ‘snap resolution’ value to the script, which dictates an amount for the boxes’ height values to round up to. The effect is a geometry that is still derived from the terrain’s topography, but has much more conviction as a potential architectural statement. The value of this is also demonstrated in a model, showing how the alignment of boxes could, in one hypothetical scenario, allow a structural system to exist across the geometry.

Party or Prison?

The theme of the camp has been interpreted as a place for temporary inhabitation, existing under a governed or imposed order. In doing so I have considered typologies such as hospitals, festivals, prisons, refugee camps, schools and military settlements. Through various spatial approaches all demonstrate a non-hierarchical approach to settlement making, with repetition of living quarters existing around common shared spaces and infrastructural frameworks.

The exception to the way this is organised in plan is the ancient city of Çatalhöyük, which has no streets or footpaths. A typology like this could be interesting to explore in section, if the role of circulation/ structure/ drainage exist at different levels to that of the living spaces. Could the camp be underground? Or elevated in the trees?


From diagram, to architecture!

IMG_1929In the final few weeks of the project I have been exploring the potential architectural and spatial qualities of the generated design. I found that when the thickness of the primary structure is fully exaggerated, the interior spaces can become dramatic and exciting. I really enjoy that the structure can now be considered with some level of architectural quality, whilst staying true to the logic of the generated form.


Carter_Fiona_Project02_Final Boards-37Carter_Fiona_Project02_Final Boards-38


Staying true to the logic. Hierarchical thickening.

Adding thickness to the structure requires careful consideration in order stay true to the geometrical logic. First, a hierarchy of member sizes was given to differentiate between the 3 systems – unit, wall, and shading panels. Instinctively these were thickened ‘perpendicularly’ creating orthogonal cross-sections.

However, I realised that to unify the geometry and avoid conflicting section joints, the cross-section must also be angular – and hence take the form of parallelograms.

Three details were modeled to demonstrate the junction types between each element of the system.

pinup sheets-06pinup sheets-07pinup sheets-09

pinup sheets-10pinup sheets-11pinup sheets-12