Robust in-line qualification of lattice structures manufactured via laser powder bed fusion

The shape complexity enabled by AM would impose new part inspection systems (e.g., x-ray computed tomography), which translate into qualification time and costs that may be not affordable. However, the layerwise nature of the process potentially allows anticipating qualification tasks in-line and in-process, leading to a quick detection of defects since their onset stage. This opportunity is particularly attractive in the presence of lattice structures, whose industrial adoption has considerably increased thanks to AM. This paper presents a novel methodology to model the quality of lattice structures at unit cell level while the part is being built, using high resolutions images of the powder bed for in-line geometry reconstruction and identification of deviations from the nominal shape. The methodology is designed to translate complex 3D shapes into 1D deviation profiles that capture the “geometrical signature” of the cell together with the reconstruction uncertainty.