Leveraging Data-Driven Design Methods to Support Complex and Multidisciplinary Design Scenarios in Additive Manufacturing

Co-chairs: Tino Stankovic, Nicholas Meisel, Serena Graziosi

Invited speakers: Chris McComb, Mark Fuge, Krishnan Suresh

The digital essence of Additive Manufacturing (AM), and thus the possibility to quite quickly generate and store a significant amount of digital data, has led to a growing interest in studying how to exploit data-driven approaches to support process planning, optimization, and monitoring in AM. The need for more reliable technologies and defect-free and repeatable processes has also driven that interest. However, examples and insights on how such approaches could be leveraged to further unlock Design for AM (DfAM) freedom are still limited. Our current attitude toward data-driven AM technology may explain the reasons behind this gap; does data just serve as a means to analyze specific situations and make informed process decisions? Or is there a greater opportunitity to leverage data-driven design methods to help us exploit and master the complex and multidisciplinary design scenarios made available through AM? What design tasks and aspects are ready for a data-driven search space exploration? What role will the DfAM expert play after the massive introduction of computational algorithms?
AM has also torn down barriers for multiscale design; hence there are significant differences in design strategies and representations among the length scales at which AM operates. It is, therefore, crucial to understand how to unify them and how ready data-driven approaches are to support this unification. The development of data-driven support technologies for AM is much more than an informed search of unconventional design solutions and alternatives within a theoretically unlimited space. It concerns tackling new design needs and finding reliable, effective, and sustainable solutions that only the design freedom allowed by AM can guarantee. Discussions are needed to clarify what data to use and how to combine them and elaborate on them. However, these discussions should also consider integrating these data into human-based intellectual processes driven by specific industrial and societal challenges. Effective data-driven solutions should be able to complement DfAM experts’ existing design thinking.
This workshop engaged participants and lecturers to reflect on several open questions. How can we as researchers improve the ability of designers to utilize the complex relationships between different length scales and oviduct abstractions (architected materials, product geometries and functions) in AM? How can we enable the practical search of the underlying solution spaces using generative design? Which aspects of the design process are ready for AI-backed support? How will the design rationale be expressed within AI algorithms to support the DfAM innovation process? How can we develop design support methods that foster human creativity in the context of their design workflows while respecting all ethical issues regarding big data? Which competencies, skills, cross-domain knowledge, and methodological support are required to enable data-driven DfAM? The workshop will offer an opportunity for design practitioners to discuss state-of-the-art methods and tools to support data-driven DfAM and their implementation in the industry. The participants will be guided towards engaging the invited lecturers using these questions and pushed to reflect on how they can contribute to data-driven DfAM through their research activities. 

The workshop took place on:

Monday, 24/Jul/2023: 

in the Workshop session 13:30-16:45

@ICED23 conference, 24 - 28 July 2023.