Original title: Influence of Extruded Filament Shape on Buildability in 3D Concrete Printing: A Geometry-Informed Deep Learning-FEM Approach
Giacomo Rizzieri, Saif-Ur-Rehman, Jörg F. Unger, Annika Robens-Radermacher · September 2026
When construction crews 3D-print concrete, a nozzle squirts out wet layers called filaments to build up walls. Engineers use computer models to test if a design will collapse during printing. Most simulations assume each printed strand is a neat, flat rectangle. In real life, wet concrete squashes and curves, making those predictions inaccurate. Researchers solved this by pairing a deep-learning tool named ShapeGen3DCP with structural simulation software. This artificial intelligence predicts true strand shapes from basic printing settings, skipping the need for slow fluid tests or physical lab trials. The team found that strand shapes strongly sway how stable walls are. True curved shapes matter most during free-flow printing, while pressed layers care less about shape simplifications. When balancing speed and accuracy, treating filaments as ellipses works well. If engineers must use simple rectangles, sizing them by conserved volume works best.
You are building a tower by stacking soft clay coils from a tube. If you assume each coil is a perfect rectangle, but it actually flattens into an oval, the tower might lean or collapse. In this paper, the researchers show that predicting the actual oval shape of extruded concrete coils makes simulations more accurate for whether the printed structure can stand up.
For engineers using 3D concrete printing, accurate predictions prevent costly structural failures. By modeling the real filament shape instead of assuming simple rectangles, simulations become more reliable, saving time and materials in construction.
The study focuses on rectilinear walls and does not address complex geometries. The framework is validated against experimental data, but broader applications or other materials are not discussed in the abstract.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Influence of Extruded Filament Shape on Buildability in 3D Concrete Printing: A Geometry-Informed Deep Learning-FEM Approach — Giacomo Rizzieri, Saif-Ur-Rehman, Jörg F. Unger et al.
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.