| Abstract Scope |
Comprehensive materials properties databases continue to be developed for additive manufacturing (AM) processes, and will eventually support the construction of design criteria, but are hindered in the near-term by the complexity of the aforesaid processes. However, with advances in data mining and curation protocols, these efforts can be readily supplemented by the evaluation of archival datasets through a meta-analysis to quantify the marginal contributions to scatter and average fatigue life with a greater level of precision and sensitivity that would otherwise be capable for most isolated studies. Not only does this approach enable a more nuanced investigation of the roles of multiple material and processing parameters but it can also identify relatively small effects. When applied to AM Ti-6Al-4V fatigue, this approach can resolve the differences between numerous datasets and enable a broader investigation of the role of processing porosity. Using a statistical model developed on archival data as input, an ensemble model was developed that helps explain the impact of pore size, shape, and location for both surface isolated and surface connected conditions. |