A novel approach for geometrical deviations determination of lattice structures from volumetric data

Additive manufacturing (AM) questions traditional approaches that manufacturer and metrologist have developed relative to formative or subtractive manufacturing technologies. As specific-to-AM parts, lattice structures are being investigated in many applications and efforts are focussed on modelling manufactured different-scale defects to predict the part properties before manufacturing. To date, X-ray computed tomography (XCT) is the most suited measuring instrument to reveal additively manufactured part internal defects, i.e. defects inaccessible to optical or tactile instruments. However, XCT requires surface determination tools which choice may be discussed. In this paper, a virtual volume correlation (V2C) method is proposed relying on modal decomposition. In other words, lattice structure shape defects are expressed directly from volumetric data, without any use of XCT surface determination tool. Extending previous works validated for representative lattice struts, entire lattice structure shape defect is identified by iteratively deforming a virtual volume, to minimise its least square grey level differences towards the measured volume. The proposed method particularly describes the strut form defects and the node displacements. This method allows to define at the same time and separately the shape defect of each strut but also the global geometrical defect of the structure. This approach is more relevant to assess the suitability of the structure with respect to mechanical loads than a deviation map that is difficult to handle.