Agent-based Methods for the Fabrication-aware Design of Modular Precast Structures from Digitally Fabricated Concrete

The construction industry is the world’s largest consumer of raw materials and emitter of CO₂, with concrete playing a central role. While concrete's fluid nature enables the creation of complex geometries, the reliance on formwork creates a paradox: to reduce manual labour and cost, conventional formwork strategies often lead to the excessive use of concrete, undermining ecological design goals. Recent developments in additive manufacturing (AM), particularly 3D-printed formworks (3DPF), open new possibilities for producing highly differentiated precast concrete elements with reduced waste and higher degrees of automation.

This doctoral research addresses the current lack of computational methods for modularising such digitally fabricated concrete (DfC) elements. The focus lies on segmenting given design geometries into discrete volumetric modules, which is a critical step that influences both the behaviour of assembled structures and respective fabrication affordances.

Modularising volumetric geometries involves navigating multiple, often competing, objectives such as module size, weight, number, and position of module interfaces. These objectives are rarely fully defined at the outset of a design process and evolve across the various planning phases. To address this, the research employs Agent-Based Modeling (ABM), where individual modules are represented as autonomous agents. This enables iterative simulation, dynamic adaptation to new constraints, and the integration of designer input into the dynamic modularisation process to ensure the applicability of the developed methods in early design phases.

The modularisation process is designed to be embedded into a Digital Twin (DT) framework via the Asset Administration Shell (AAS), aligning with Industry 4.0 standards. This enables the potential integration with cyber-physical production systems (CPPS) and supports a seamless transition from design to automated fabrication.

The developed methods aim to enhance the efficiency, sustainability, and possibility to design and build with highly differentiated building elements in precast concrete, addressing ongoing challenges in the contemporary construction industry.

ICD Institute for Computational Design and Construction, University of Stuttgart

David Stieler, Prof. A. Menges

DFG Priority Program SPP 2187, Project TP 09

German Research Foundation (DFG), EXC 2120 IntCDC

Kolbeck, L., Kovaleva, D., Manny, A., Stieler, D., Rettinger, M., Renz, R., Tošić, Z., Teschemacher, T., Stindt, J., Forman, P., Borrmann, A., Blandini, L., Stempniewski, L., Stark, A., Menges, A., Schlaich, M., Albers, A., Lordick, D., Bletzinger, K.-U., & Mark, P. (2023). Modularisation Strategies for Individualised Precast Construction—Conceptual Fundamentals and Research Directions. Designs, 7, Article 6. https://www.mdpi.com/2411-9660/7/6/143

Stieler, D., Schwinn, T., & Menges, A. (2022). Volumetric intersections: Modularization approaches for freeform prefab concrete construction. Civil Engineering Design, 4, Article 1–3. https://doi.org/10.1002/cend.202100047

Stieler, D., Schwinn, T., & Menges, A. (2022). Automatisierte Bauteilzerlegung für Betonfertigteile aus additiv hergestellten Schalungen. Beton- Und Stahlbetonbau, 117, Article 5. https://doi.org/10.1002/best.202200006

Forman, P., Stieler, D., Mark, P., & Menges, A. (2022). Adaptive Modulbauwesien mit Methoden der Fließfertigung. In V. T. Nguyen, M. Krüger, B. Freytag, & T. M. Laggner (Eds.), 5. Grazer Betonkolloquium (pp. 79–86). Verlag der TU Graz.

Stieler, D., Schwinn, T., & Menges, A. (2022). Additive formwork in precast construction - Agent-based Methods for Fabrication-aware Modularization of Concrete Building Elements. Post Carbon - Proceedings of the 26th International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2022. https://caadria2022.org/wp-content/uploads/2022/04/435-1.pdf

Stieler, D., Schwinn, T., Leder, S., Maierhofer, M., Kannenberg, F., & Menges, A. (2022). Agent-based modeling and simulation in architecture. Automation in Construction, 141, 104426. https://doi.org/10.1016/j.autcon.2022.104426