Times Eureka Pavilion (eng)
NEX - Alan Dempsey, Paul Loh, Michal Piasecki, Tomasz Starczewski, James Chung
ENG - Designed in collaboration with Landscape Designer Marcus Barnett, the temporary garden and pavilion explore the significance of plants to science and society. Plant species were chosen to reflect their benefits to society including medicinal, commercial and industrial uses found in everyday life.
The design brief for the garden was extended to the pavilion by looking closely at the cellular structure of plants and their processes of growth and formation. The final structure was designed using computer algorithms that mimic natural growth and was intended to allow visitors to experience the patterns of biological structure at an unfamiliar scale.
The walls and roof of the pavilion were designed as a deep structure that defines a spatial enclosure, but at the same time remains porous to filter light, the movement of air and sound, and to frame and direct views of the garden beyond. The geometry of all of the cassettes and cells radiates from two central vanishing points to create a dramatic and changing visual experience of the garden as visitors move through the structure.
Manufactured from sustainably sourced timber, the primary timber ‘capillaries’ form the supporting structure, which are inset with secondary timber cassettes that hold an infill of translucent recycled plastic cells. The cassettes in the roof are sealed with laminated glass. In operation the pavilion continues to mimic water transfer found in plant biology. Rain water literally runs off the glazed roof cells into the main recessed capillaries and down the walls to the ground.
Times Eureka Pavilion Design process
The design of the project explored the use of recursion as a generative process to evolve a 3 dimensional structure that could be occupied by people. The technique of recursion is the process of repeating actions or objects in a self-similar way in successive steps. The development of the system is governed by a series of rules that define local structure without determining overall form in advance. Such mathematical growth processes offer an interesting alignment between computation and nature by producing unique and non-repeating instances of form that obey consistent rules of formation.
We used two recursive approaches In the Times Eureka pavilion. The first experiments used a Lindenmeyer System (L-systems) algorithm to generate a two dimensional pattern that was then wrapped over a 3D bounding box of the pavilion envelope. Versions of this were iterated many times using different values of input parameters (initial length of the branches, number of branches branching out in every generation, range of angles between branches). While this exploration revealed a basic arrangement of the primary structure, we realised that it tended to fix the pattern at a particular scale that made the branching appear too literal i.e. there was a direct correlation between the scale of the branching in the pavilion and the scale of the branching in the trees in the surrounding garden. This was something we wanted to avoid, as we felt it would be more interesting for users to experience the patterns of plant biology at an unfamiliar scale.
To make this shift we switched to a process of decomposing the spaces between the main structure using voronoi procedures, much like the way cell division occurs in plants. The process of division was recursive at three scales; the first was the primary structure of 140mm wide beams; the second was at the scale of 20mm walled ‘cassettes’, of which there were 136 unique elements. The final scale was the cell infill made from recycled translucent polypropylene. Again these were individually unique and there were 586 in total.
Once assembled on site the quality of the patterning across the walls, floor and ceiling of the structure produced a powerful sense of occupying a biological structure. In operation the pavilion continued to mimic water transfer in plant biology. Rain water literally runs of the roof infill cells into the main capillaries in the walls of the cube from the roof into the ground.
61 users love this project