Direct additive manufactured beam shape defect identification from computed tomography and modal decomposition
| Autor: | Marc-Antoine de Pastre, Yann Quinsat, Claire Lartigue |
|---|---|
| Přispěvatelé: | Laboratoire Universitaire de Recherche en Production Automatisée (LURPA), Université Paris-Saclay-Ecole Normale Supérieure Paris-Saclay (ENS Paris Saclay) |
| Jazyk: | angličtina |
| Rok vydání: | 2021 |
| Předmět: | |
| Zdroj: | Joint Special Interest Group meeting between euspen and ASPE Advancing Precision in Additive Manufacturing Joint Special Interest Group meeting between euspen and ASPE Advancing Precision in Additive Manufacturing, Sep 2021, St Gallen, Switzerland HAL |
| Popis: | International audience; As one of the main additive manufacturing (AM) advantages, lattice structures are being studied in many applications such as vibration attenuation, weight reduction of components or optimised heat exchangers. However, lattice structures are challenging to produce, and may present some shape defects. Although significant works have been performed in lattice structure defect observations such as overhanging features or resulting porosity, there has been relatively less research in modelling shape defects by defining a geometric description approach. In this paper, a Virtual Volume Correlation (V2C) method is proposed in order to identify metal laser powder bed fusion (LPBF) BCCz struts shape defect directly from volumetric data obtained by X-ray computed tomography (XCT). In the proposed V2C method, a correlation score is calculated between the volumetric data and a virtual volume. This virtual volume is determined according to the computer-aided-design (CAD) model and a shape defect which is defined using a linear decomposition relying on a user-defined defect basis. Shape defects of the studied part are successively, according to a Newton Raphson optimisation scheme, determined by correlation score minimisation. Vertical and inclined beams have been printed and measured with XCT and focus variation (FV). Strut geometries obtained with V2C methodology are compared with extracted ISO50% point clouds, on the one hand, and measured FV point clouds, on the other hand, by computing signed cloud-to-mesh distances. These comparisons bring out that the V2C method is efficient to identify strut shape defects directly from volumetric data, without any post-reconstruction XCT data treatment. The simplification of these data treatment steps then raises the direct and accurate CAD feedback opportunity. Conclusions are drawn regarding the suitability of the proposed V2C method and its further development to more complex LPBF structures. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|