Pore-affected fatigue life scattering and prediction of additively manufactured Inconel 718: An investigation based on miniature specimen testing and machine learning approach

• Published in: Materials science & engineering. A, Structural materials : properties, microstructure and processing Structural materials : properties, microstructure and processing, 2021-01, Vol.802, p.140693, Article 140693
• Main Authors: Luo, Y.W., Zhang, B., Feng, X., Song, Z.M., Qi, X.B., Li, C.P., Chen, G.F., Zhang, G.P.
• Format: Article
• Language: English
• Subjects: Algorithms; Fatigue life; Fatigue tests; Laser beam melting; Machine learning; Nickel base alloys; Pore feature; Porosity; Scattering; Selective laser melting; Statistical analysis; Superalloys
• ISSN: 0921-5093; 1873-4936
• DOI: 10.1016/j.msea.2020.140693

Fatigue life scattering and prediction of Inconel 718 fabricated by selective laser melting were investigated using miniature specimen tests combined with statistical method and machine learning algorithms. The relationship between pore features and fatigue life of the selective laser melting-fabricated specimens was analyzed statistically. The results show that the increase in the size and/or the number of the pores in the specimens, and/or the decrease in the distance from a pore center to the specimen surface degraded the fatigue life. The machine learning and statistical analysis results reveal that the fatigue life are most closely related to the location of the pores compared with the size and the number of pores in the specimens. The finding may provide a potential way to get high-throughput statistical data helping in evaluating defect-dominated scattering and prediction of fatigue life of additive manufactured metallic parts using miniature specimen testing assisted by the machine learning approach.