Page 13 - ESAM-1-1

Engineering Science in
Additive Manufacturing ML in MAM monitoring and control through images

A

B

Figure 5. Schematic diagram of an X-ray imaging system: (A) Laser powder bed fusion.43 Reproduced with permission from Elsevier. Copyright © 2018
The Authors; (B) direct energy deposition. Reproduced with permission from Elsevier. Copyright © 2023 The Authors. Both are distributed under a
62
CC-BY 4.0 license (https://creativecommons.org/licenses/by/4.0/).
and structural changes within the sample. For example, addition, Ito et al. developed an acoustic sensing-based
66
Tempelman et al. and Li et al. 63,64 recorded acoustic signals method to detect microdefects, such as cracks and pores,
produced during the LPBF process to detect the onset of at various locations in the LPBF process.
internal pores. They then successfully extracted feature
vectors that correlated with internal pore formation. Acoustic imaging proves to be a valuable tool in
In the DED process, Chen et al. recorded raw acoustic MAM for detecting defects, predicting pore formation,
65
signals and marked regions where defects were identified. and enhancing the overall quality control of MAM
Subsequently, they trained a CNN model to predict processes. Its ability to capture real-time acoustic signals
the occurrence of keyhole pores with 93% accuracy. In enables researchers to have a better understanding of

Volume 1 Issue 1 (2025) 7 doi: 10.36922/esam.8548