Learning From Physical Data
In this phase of the project the possibilities of working with physical design artifacts were investigated. The experiment was aiming to learn from the designers tacit knowledge by analyzing physical models and their performances created by the designers, to design new artifacts using artificial intelligence that base on the extracted rules and patterns from physical environment.
The experiment was conducted during an artist in residency month in September 2017 in IaaC Barcelona in collaboration with the tutors of Open Thesis Fabrication(OTF) Program Edouard Cabay, Alexandre Dubor, Kunal Chandra and the students of this program.
The students were asked to design ‘Zeer Pots’, which is a pot in pot cooling system basing on the principle evaporation. This kind of design problems are hard to simulate digitally, since the design performance depend on the complex relations of certain morphological features, material properties, environmental properties. To understand the complex behavior a design, fabricate and analyze workflow should be followed. This would be a very long process and iterative analysis process. However, by using artificial intelligence we were able to create a short cut, that can predict the behavior of certain geometric features digitally.