Compressive capacity of concrete dry joints across shotcrete, extrusion, and particle-bed additive manufacturing with robotic subtractive finishing

Abstract Additively manufactured (AM) concrete segments often require reliable, repeatable dry-joint interfaces, yet segmentation quality and dimensional accuracy remain challenging. This study evaluates dry joints produced by robotic subtractive finishing (sawing/milling) as a practical interface strategy for segmented AM concrete construction. Four joint geometries (Smooth, Triangular, Arc, and Trun-Pyramid) were fabricated as interlocking pairs from blanks produced by three AM routes (shotcrete, extrusion, and particle-bed printing). Specimens with build orientations of $$0^\circ$$ 0 ∘ and/or $$90^\circ$$ 90 ∘ were tested under monotonic uniaxial compression. Cross-method comparability was supported by batch-matched companion-cube tests, laser-scan-to-CAD geometric assessments, and process measurements, including recorded machining durations for milled/sawn configurations. In total, 120 tests were performed (102 monolithic/jointed compression tests and 18 companion-cube tests). Several joint configurations reached compressive capacities close to the corresponding monolithic references. The Arc geometry showed the most consistent high performance across all AM routes. For shotcrete printing, the sawn Triangular joint provided the most favorable performance-to-machining-time trade-off, indicating substantial reductions in finishing effort. Joint geometry, AM route, build orientation, and subtractive finishing jointly govern both compressive capacity and fabrication efficiency. The reported dataset provides quantitative guidance for selecting dry-joint details in segmented AM concrete elements.