Virtual volume correlation of lattice structures: from volumetric data to geometrical and dimensional defects identification

Additively manufactured lattice structures are widely studied in the literature for their advantageous properties. Manufacturing steps introduce defects that are often classified amongst geometry and dimension, surface quality, and porosity defects. Because these defects drastically affect lattice structures' behaviour, the manufactured parts might lack their required properties. Assessments of these properties are mainly performed by finite-element analysis. This approach primarily relies on meshed X-ray computed tomography (XCT) measurements or meshed computer-aided design (CAD) models and less on parametric models where defects can be easily implemented. Although the former is the most representative of the real part geometry, measured geometrical features are implicit in this analysis and cannot be identified in a direct parametric approach. This paper proposes a virtual volume correlation (V2C) methodology to extract geometrical and dimensional defects of laser powder bed fusion lattice structures. The V2C methodology only relies on volumetric measurement and does not require additional XCT post-reconstruction steps. Such a methodology combines strut and node displacement fields that are decomposed as a sum of chosen parametric elementary displacement fields. V2C is first validated on generated volumes where known defects are introduced, and its ability to relocate these introduced defects is evaluated. Relocation errors are found to be lower than half a voxel. Then, V2C is applied to manufactured lattice structures that differ in position and orientation. Results show the dependence of strut typology, whether vertical or inclined strut, on the resulting geometrical and dimensional defects. V2C also shows the orientation and the position effect of the lattice structure's geometrical and dimensional defect within the building chamber where lattice nodes' displacement amplitudes are about 30 µm. The conclusion outlines the parametric identification of geometrical and dimensional defects that can be easily implemented for further finite element models.