A new FE post-processor for probabilistic fatigue assessment in the presence of defects and its application to AM parts

Abstract

Despite the disruptive benefits of Additive Manufacturing (AM), the application of this technology for safety-critical structural parts in aerospace is still far from being achieved and standardised. The necessity to comply with very strict reliability requirements is hindering this final step because of the large scatter and low reproducibility always associated with AM, especially in terms of fatigue strength. In this regard, manufacturing defects are the most important and complex issue, but several other sources of variability have an effect as well. The AM community and the main aerospace industries involved are starting to agree that damage-tolerant approaches are necessary and that probabilistic methods are best-suited to obtain reliable but not over-constrained assessments. To address this issue, the authors have developed Pro-FACE, a fully-probabilistic software that aims to robustly assess the fatigue strength and critical locations of complex components in the presence of defects. This paper presents the underlying concept, its implementation and early validation, and a simple application to a space component. The analytical formulation makes this tool ideal to evaluate very low failure probabilities with limited time and effort, which can provide valuable information to significantly improve part design and qualification.

A numerical framework to predict the fatigue life of lattice structures built by additive manufacturing

“…11b shows large stress concentrated areas around nodes. In fact, strut failure results from combined effects of stress concentration and surface notch-like defects, as recently discussed by Persenot et al [7] or Romano et al [45].…”

Killer notches: The effect of as-built surface roughness on fatigue failure in AlSi10Mg produced by laser powder bed fusion

“…It is interesting to observe that, due to bending of the specimens, the failures effectively started from the surface features and from this point of view, it looks that the assumption by Molaei et al [40] for axial specimens that life prediction could be made considering the sum of surface plus maximum internal defect sizes is excessively conservative in the case of a stress gradient (this is due to the fact that the real probability of occurrence near the surface has to be considered [67]).…”

Defects in Electron Beam Melted Ti-6Al-4V: Fatigue Life Prediction Using Experimental Data and Extreme Value Statistics

“…An alternative to using the Murakami criterion could be to consider all defects within a fixed sub surface volume to be near the surface and thus included in the analysis. This would require determination of such a volume whilst accounting for the geometry of the defects present in the specific material, which can be challenging [ 42 ]. Another solution that avoids the need to classify defects as near the surface would be to consider a different (internal) crack case.…”