During the design process, much more information should be available and visualizable in real time, Photo: © ETH Zurich

When architects and engineers plan structures and develop them technically, it usually takes a lot of time to get from the initial draft to the first sod being turned. Various blueprints are discussed step by step in terms of cost, engineering, the materials used, equipment and scheduling during a series of different meetings while weighing up the benefits and disadvantages. Digital design methods, including ones that use artificial intelligence (AI) and extended reality (XR), are now poised to change this process from the ground up. ETH Zurich is opening a new Centre for Augmented Computational Design in Architecture, Engineering and Construction, with an aim of bringing together the work of architects and engineers in a close, interdisciplinary way. Named Design++, it sees 22 professors joining forces from the Architecture, Civil, Environmental and Geomatic Engineering, Computer Science, Materials, and Mechanical and Process Engineering departments and the National Centre of Competence in Research Digital Fabrication (NCCR DFAB). Moreover, a new position for Professor of Computational Augmented Design has been created and will be advertised shortly.

Bridge building is an area where methods such as these can change the traditional model. Bridges are highly complex structures. However, realistic structural analysis (modelling all the relevant forces mathematically with various influences) is often too great an effort during the early stages of a design. Consequently, bridge builders devise structural concepts without optimising material consumption or ecological factors, for instance. The result of this is that relatively large reserves need to be integrated into them. Walter Kaufmann, Chair of Structural Engineering at ETH Zurich, and his team are seeking to change this way of doing things. They are performing digital structure analyses for bridges defined by parameters, with these analyses being used to train artificial intelligence. AI is therefore intended to be able to accomplish such structural analysis in substantially shorter times. Using specified bridge parameters, AI will ultimately even be able to generate rough designs itself, to which finer detail will be added in interaction with the designer. Today it often takes months until bridge builders can present their initial designs. With this new method, it will take them just days or hours.

Another research project involves the development of augmented computational tools that can be used to explore creative leeway in a better way. Instead of laboriously adapting a design until all the criteria and objectives are met, AI models are planned to suggest different design variants. The tools are being tested on pillars and beams, transparent facades and special wall elements for indoor acoustics. Robert Flatt, Professor of Physical Chemistry and one of the originators of the idea behind Design++, says, “We are striving to expand creativity. We are looking to optimise buildings using new design tools while digitally expanding the scope for design.”

Share this page on Social Media:

Print Friendly, PDF & Email