We present a machine learning workflow to discover signatures in acoustic measurements that can be utilized to create a low-dimensional model to accurately predict the location of keyhole pores formed during additive manufacturing processes. Acoustic measurements were sampled at 100 kHz during single-layer laser powder bed fusion (LPBF) experiments, and spatio-temporal registration of pore locations was obtained from post-build radiography. Power spectral density (PSD) estimates of the acoustic data were then decomposed using non-negative matrix factorization with custom
This content will become publicly available on December 1, 2024
- Award ID(s):
- 1905910
- NSF-PAR ID:
- 10464092
- Date Published:
- Journal Name:
- Communications Materials
- Volume:
- 4
- Issue:
- 1
- ISSN:
- 2662-4443
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract $$\varvec{k}$$ $$\varvec{k}$$ $$\varvec{k}$$ blindly construct a feature space based on the PSD estimates in anoptimized fashion. Moreover, the NMF$$\varvec{k}$$ k , the relatively light computational cost of the machine learning workflow, and the high classification accuracy makes the proposed workflow an attractive candidate for edge computing toward in-situ keyhole pore prediction in LPBF. -
Additively manufactured electronics (AMEs), also known as printed electronics, are becoming increasingly important for the anticipated Internet of Things (IoT). This requires manufacturing technologies that allow the integration of various pure functional materials and devices onto different flexible and rigid surfaces. However, the current ink-based technologies suffer from complex and expensive ink formulation, ink-associated contaminations (additives/solvents), and limited sources of printing materials. Thus, printing contamination-free and multimaterial structures and devices is challenging. Here, a multimaterial additive nanomanufacturing (M-ANM) technique utilizing directed laser deposition at the nano and microscale is demonstrated, allowing the printing of lateral and vertical hybrid structures and devices. This M-ANM technique involves pulsed laser ablation of solid targets placed on a target carousel inside the printer head for in-situ generation of contamination-free nanoparticles, which are then guided via a carrier gas toward the nozzle and onto the surface of the substrate, where they are sintered and printed in real-time by a second laser. The target carousel brings a particular target in engagement with the ablation laser beam in predetermined sequences to print multiple materials, including metals, semiconductors, and insulators, in a single process. Using this M-ANM technique, various multimaterial devices such as silver/zinc oxide (Ag/ZnO) photodetector and hybrid silver/aluminum oxide (Ag/Al2O3) circuits are printed and characterized. The quality and versatility of our M-ANM technique offer a potential manufacturing option for emerging IoT.more » « less
-
more » « less
Abstract Laser processing of thermoelectric materials provides an avenue to influence the nano‐ and micro‐structure of the material and enable additive manufacturing processes that facilitate freeform device shapes, a capability that is lacking in thermoelectric materials processing. This paper describes the multiscale structures formed in selenium‐doped bismuth telluride, an n‐type thermoelectric material, from laser‐induced rapid melting and solidification. Macroscale samples are fabricated in a layer‐by‐layer technique using laser powder bed fusion (also known as selective laser melting). Laser processing results in highly textured columnar grains oriented in the build direction, nanoscale inclusions, and a shift in the primary charge carriers. Sparse oxide inclusions and tellurium segregation shift the material to p‐type behavior with a Seebeck coefficient that peaks at 143 µV K–1at 95 °C. With an average relative density of 74%, fabricated parts have multiscale porosity and microscale cracking that likely resulted from low powder layer packing density and processing parameters near the transition threshold between conduction and keyhole mode processing. These results provide insights regarding the pathways for influencing carrier transport in thermoelectric materials via laser melting‐induced nanoscale structuring and the laser processing parameters required to achieve effective powder consolidation and hierarchical structuring in thermoelectric parts.
-
Physics-informed machine learning is emerging through vast methodologies and in various applications. This paper discovers physics-based custom loss functions as an implementable solution to additive manufacturing (AM). Specifically, laser metal deposition (LMD) is an AM process where a laser beam melts deposited powder, and the dissolved particles fuse to produce metal components. Porosity, or small cavities that form in this printed structure, is generally considered one of the most destructive defects in metal AM. Traditionally, computer tomography scans measure porosity. While this is useful for understanding the nature of pore formation and its characteristics, purely physics-driven models lack real-time prediction ability. Meanwhile, a purely deep learning approach to porosity prediction leaves valuable physics knowledge behind. In this paper, a hybrid model that uses both empirical and simulated LMD data is created to show how various physics-informed loss functions impact the accuracy, precision, and recall of a baseline deep learning model for porosity prediction. In particular, some versions of the physics-informed model can improve the precision of the baseline deep learning-only model (albeit at the expense of overall accuracy).more » « less
-
more » « less
Abstract Dynamic solidification behavior during metal additive manufacturing directly influences the as-built microstructure, defects, and mechanical properties of printed parts. How the formation of these features is driven by temperature variation (e.g., thermal gradient magnitude and solidification front velocity) has been studied extensively in metal additive manufacturing, with synchrotron x-ray imaging becoming a critical tool to monitor these processes. Here, we extend these efforts to monitoring full thermomechanical deformation during solidification through the use of operando x-ray diffraction during laser melting. With operando diffraction, we analyze thermomechanical deformation modes such as torsion, bending, fragmentation, assimilation, oscillation, and interdendritic growth. Understanding such phenomena can aid the optimization of printing strategies to obtain specific microstructural features, including localized misorientations, dislocation substructure, and grain boundary character. The interpretation of operando diffraction results is supported by post-mortem electron backscatter diffraction analyses.
