Benchmark of a probabilistic fatigue software based on machined and as-built components manufactured in AlSi10Mg by L-PBF. (December 2022)

• Record Type: Journal Article
• Authors: Sausto, F.; Romano, S.; Patriarca, L.; Miccoli, S.; Beretta, S.
• Is Part Of: International journal of fatigue. Volume 165(2022)
• Publication Date: 2022-12
• Subjects: Additive manufacturing -- Fatigue -- Defects -- As-built surface -- Failure probability -- Residual stresses; Materials -- Fatigue -- Periodicals; Materials -- Fatigue; Periodicals; 620.1122
• DOI: 10.1016/j.ijfatigue.2022.107171
• Languages: English
• ISSNs: 0142-1123

Abstract

Abstract: The possibility to obtain optimized components with a reduced weight is the main driver of space and aeronautic industries in seriously considering the metal additive manufacturing (AM) technology for production. Despite the incontrovertible advantages offered by this manufacturing technique, the material produced is usually affected by the presence of internal defects, a poor surface quality, and process-induced residual stresses. These features strongly affect the fatigue performance and reproducibility of AMed parts, limiting the adoption of deterministic criteria for fatigue assessment. A probabilistic approach is therefore needed for the analysis of critical and structural components. To this aim, a fully probabilistic finite element (FE) post-processor, ProFACE, was developed by part of the authors to assess the fatigue strength and critical locations of complex components in the presence of process-induced defects. A wide benchmark activity was supported by the European Space Agency (ESA) to test the software capabilities for the life prediction of components manufactured in AlSi10Mg by L-PBF. After tuning ProFACE parameters based on the results obtained on standard fatigue specimens, the software was used to estimate the fatigue life of the components obtaining a good description of the experimental dataset for both volumetric and surface defects. The software was then used to explore the effect of the variability of the most significant parameters affecting fatigue strength of AlSi10Mg AMed components. Highlights: Fatigue performances of AMed parts are strongly influenced by internal defects, surface features and residual stresses. To handle these random features a probabilistic fatigue postprocessor was developed. The main software's features were tested and compared to S-N curves of benchmark components. The failure probability of AMed components was correctly estimated by the software. Software outputs showed to be a suitable tool for the qualification of AMed parts.