Statistical approach to quantifying the spatial distribution of porosity defects in Inconel 718, A

Abstract

As 3D imaging of materials becomes increasingly accessible the materials science community has a growing need for relevant statistics that quantify the internal structure of manufactured parts. Such statistics can be sensitive to the orientation of the imaged part; we implement a method of orienting imaging data by fitting by least squares simple geometric surfaces to 3D data. A Monte-Carlo based method for evaluating the statistical uncertainty associated with such orientation is also proposed. An application of this model is explored using samples of Inconel 718 built through additive manufacturing. A subsequent analysis attempts to detect layers in the porosity defect structure of Inconel 718 samples. The analysis found little evidence for layering.