papersSEP 10 04:00 UTC
Deep Learning-FEM Approach Links Extruded Filament Shape to Buildability in 3D Concrete Printing
Researchers present a combined deep learning and finite element framework that factors in the real cross-sectional geometry of extruded concrete filaments when evaluating whether printed layers can bear the weight of subsequent ones. The work addresses a common limitation of existing buildability assessments, which typically rely on simplified filament shapes, potentially misjudging the stability of 3D-printed concrete structures.