Leveraging the digital thread for physics-based prediction of microstructure heterogeneity in additively manufactured parts

Journal Article · 07 November 2023 · Additive Manufacturing

DOI: https://doi.org/10.1016/j.addma.2023.103861 · OSTI ID:2224202

Knapp, Gerald L.  ^[1]; Stump, Benjamin ^[1]; Scime, Luke  ^[1]; Márquez Rossy, Andrés ^[1]; Joslin, Chase ^[1]; Halsey, William  ^[1]; Plotkowski, Alex  ^[1]

---

1. Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)

A major limitation of additive manufacturing (AM) processes is that local conditions of material deposition frequently lead to unintentional heterogeneities in microstructure and properties within a single component, despite nominally uniform process conditions. Up to now, there has been no way to a priori determine the distribution of these heterogeneities, requiring expensive trial-and-error approaches to fabrication, testing, and characterization. Here, a physics-based framework for creating a digital representation of the laser powder bed fusion (PBF) process is proposed to predict the variation in solidification behavior that leads to heterogeneous microstructures in an as-built part. By leveraging in situ process data stored in the part’s digital thread, the scan path and process parameters were input into a heat transfer model which predicted solidification data at the melt pool scale. A two-step unsupervised clustering algorithm was used to first cluster the local solidification conditions (12.5µm³ voxels) and then to cluster the regional behavior on the scale of multiple scan passes and print layers (250µm³ super-voxels). This process was used to identify regions with similar solidification characteristics for multiple locations in a Stainless Steel 316-L component. The corresponding as-built part was sectioned and characterized using electron backscatter diffraction (EBSD). Quantitative analysis of the pole figures confirmed that the predicted regions of heterogeneity in the solidification conditions corresponded with differences in the observed microstructure. In conclusion, this work shows a viable path for estimating the microstructural heterogeneity for additively manufactured parts to either limit microstructural variation throughout a part or to enable functionality-based variation of the microstructure.

• Research Organization: Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States). Manufacturing Demonstration Facility (MDF); Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States). Compute and Data Environment for Science (CADES)
• Sponsoring Organization: USDOE Office of Energy Efficiency and Renewable Energy (EERE), Energy Efficiency Office. Advanced Materials & Manufacturing Technologies Office (AMMTO); USDOE Office of Nuclear Energy (NE)
• Grant/Contract Number: AC05-00OR22725
• OSTI ID: 2224202
• Journal Information: Additive Manufacturing, Vol. 78; ISSN 2214-8604
• Publisher: ElsevierCopyright Statement
• Country of Publication: United States
• Language: English