Computer-Aided Fabrication Technologies as Computational Design Mediators

Sönmez A., SORGUÇ A.

39th International Conference on Education and Research in Computer Aided Architectural Design in Europe, eCAADe 2021, Novi-Sad, Serbia, 8 - 10 September 2021, vol.1, pp.465-474 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 1
  • City: Novi-Sad
  • Country: Serbia
  • Page Numbers: pp.465-474
  • Keywords: Arcihtecture 4.0, Artificial Intelligence, Computer-Aided Fabrication, Digital Tectonics, Generative Design
  • Middle East Technical University Affiliated: Yes


© 2021, Education and research in Computer Aided Architectural Design in Europe. All rights reserved.The developments in recent technologies through Industry 4.0 lead to the integration of digital design and manufacturing processes. Albeit manufacturing continues to increase its importance as design input, it is generally considered at the last stages of the design process. This misconception results in a gap between digital design and fabrication, leading to differences between the initial design and the fabricated outcome in the context of architectural tectonics. Here, we present an artificial intelligence (AI)-based approach that aims to provide a basis to bridge the gap between computation and fabrication. We considered a case study of a 3D model in two stages. In the first stage, an intuitive and top-down design process is adopted, and in the second stage, an AI-based exploration is conducted with three cases derived from the same 3D model. The outcomes of the two stages provided a dataset including different design parameters to be used in a decision tree classifier algorithm which selects the manufacturing method for a given 3D model. Our results show that generative design simulations based on manufacturing constraints can provide a significant variety of manufacturable design alternatives, and minimizes the difference between design alternatives. Using our proposed approach, the time spent in form-finding and fabrication can be reduced significantly. Additionally, the implementation of decision tree classifier learning algorithm shows that AI can serve designers to make accurate predictions for manufacturing method.