Computational Construction

Computational Construction

Design result by Arjun Sharma

Design result by Arjun Sharma

Digitally generated and machine-fabricated components expand the construction spectrum and incorporate the factors associated with materials decisions and manufacturing issues into the design process. Insights from the manufacturing process are fed back into the drafting process as programming parameters. These production methods, however, will not necessarily result in formal transfor­mations in architecture. Instead, it is a question of applying methods in order to further refine and optimize processes. To this extent, digital fabrication methods are capable of more than simply creating individual structures with a higher degree of complexity.

Against this background the Computational Construction module aimed at an integration of production and construction parameters into the further design process. The findings from the winter term module Computational Design were the starting point for the development of the individual realization strategy for each project. Hence, the ‘tectonic’ idea was continued towards a prototypical structure, that is informed by its materiality, production and assembly as well as structural and architectural performance.

3d-model 1

3d-model by Arjun Sharma

The project of Arjun Sharma focused on computational strategy for the creation of minimal surfaces and their materialization. As a first step the shape of a minimal surface is simulated using digital processes. These processes allow for the incorporation of the materials properties (stretchability), early in the design stage. The Ideal tension to be applied to the fabric is calculated. The constructions details for the supporting frameworks are also integrated. The cutting patterns for the fabric are prepared after considering all these factors. After carrying out various material experiments, scale models and prototypes a 1:1 construction was realized on the Campus in cooperation with Ellermann Objektbau (www.ellermann-objektbau.de).

arjun-01

3d-model and unrolled surfaces

model making

Model making

arjun-03

Paper model of the design

Ahmed Moharam developed a stress-based structural design using the Voronoi pattern as a tessellation principle for a double curved sheltering roof. After the initial form-finding and shape optimization using a thrust network analysis (TNA) the stress-lines were translated into a triangular grid layout from which the dual pattern of the Voronoi resulted. For the construction and materialization of the roof structure a scale model was built in cardboard to verify the computational concept and a series of 1:1 prototypes of the cells were realized in Alucubond.

ahmed_01

Merged image of wireframe model and photography

ahmed_03

Parametric model of the structure

ahmed_02

Subdivision of the guiding surface

ahmed_05

Final model made of lasercutted cardboard

The Gothic roof structure of Westminster Abbey Church in London served as a reference for the case study by Alexander Fillies. Taking the principle of rotational lines that form a three-dimensional shape as a starting point, the project focused in a first step on the generation of a computational model that allowed for the simulation of ruled rotational surfaces, more specifically hyperboloids of one sheet, e.g. they contain at least one family of straight lines. For the materialization a series of material tests (metal wire, rubber cord, fiberglass roving) were carried out to investigate the structural performance, both in tension and compression. Based on this research Alexander chose for a fiberglass structure as a construction strategy. The fiberglass wires were woven around an adjustable formwork and reinforced by epoxy resin and hardener. The resulting knot structure proofed to be extremely light and at the same time very stable.

alex_01

Final result by Alexander Fillies

alex_05

Supporting construction for model creation

alex_04

Digital model and parametric definition in Rhinoceros 3D

alex_03

Possible variations defined in the parametric definition

alex_02

Amount of subdivisions and resolution of the result also defined in the parametric definition

alex_06

Final model by alexander Fillies

Students: Arjun Sharma, Ahmed Moharam, Alexander Fillies, Suraj Sunil Kumar Shetty

Lecturers: Prof. Marco Hemmerling, Matthias Michel