Comparison of different approaches to model fatigue for additively manufactured specimens considering production-related characteristics

Abstract

Additive manufacturing processes such as selective laser melting (SLM) gain more and more importance and are already applied in various industries ranging e.g. from aerospace to automotive. However, in order to further spread their distribution among the common manufacturing technologies, reliable material data is required. While static material data for various materials is available in literature, fatigue data is comparably scarce. The properties of a specimen vary depending on the building direction and the building parameters. Furthermore, due to the manufacturing process, the parts show especially high surface roughness, isolated defects at overall low porosity and significant macroscopic residual stresses in the unprocessed state. As it is not always possible to alter these features, it seems advisable to incorporate them in a fatigue model for SLM specimens, especially as these are the decisive features regarding fatigue strength. Estimation of the fatigue limit is carried out via the defect stress gradient (DSG) approach and a Weibull distribution. Both approaches are capable to include the aforementioned features and are therefore used and compared for the fatigue strength calculations of 316L produced via SLM.