Geometry-based prediction of built imperfections for additively manufactured complex metamaterials

Despite the high structural flexibility offered by additive manufacturing, the effective manufacturability of complex metamaterials with highly non-periodic architectures is often limited by process-induced defects such as unsupported overhangs, disconnected features, and enclosed cavities. These limitations are particularly critical for disordered structures, where geometric complexity precludes straightforward design rules and systematic screening. We present an algorithmic framework that relies solely on geometric criteria to detect critical features in laser-based powder bed fusion printing. The method, based on persistent homology and local geometric criteria, identifies regions prone to fabrication failure through voxel-based analysis, including unsupported material, restricted-access voids, and weak connections. These criteria are combined into a unified procedure to assess the printability of candidate structures prior to manufacturing. The approach is demonstrated on spinodoid metamaterials, whose morphologies provide an ideal test case for complex and/or disordered structures. To experimentally validate the proposed methodology, comparisons with several samples produced by laser-based powder bed fusion of metals (PBF-LB/M) using 316L powder are presented. Overall, we see that all critical defects were correctly isolated from other structural arrangements. The presented framework, which can be extended to other structures and additive manufacturing techniques, enables efficient, process-aware screening of complex architectures and can be applied to complement the design of other metamaterials incorporating manufacturability.