Machine learning–aided real-time detection of keyhole pore generation in laser powder bed fusion

Ren, Zhongshu; Gao, Lin; Clark, Samuel J.; Fezzaa, Kamel; Shevchenko, Pavel; Choi, Ann; Everhart, Wes; Rollett, Anthony D.; Chen, Lianyi; Sun, Tao

doi:10.1126/science.add4667

Cited by 74 publications

(23 citation statements)

References 54 publications

(46 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[9][10][11][12] These process parameters can be studied individually or using some synthetic process parameter indexes such as energy density, such as linear energy density (P/v), area energy density (P/(vd)), volumetric laser energy density (P/(vdt)). [13][14][15] The results show there is proper parameters or laser energy density. Zhang et al, 15 showed that the threshold laser energy density for the Al-Cu-Mg powder is 340 J/mm 3 .…”

Section: Introductionmentioning

confidence: 93%

See 1 more Smart Citation

3D characterization of the microstructure of LPBF- fabricated Inconel 718 alloy

Kang¹,

Huangb²,

Yub³

2023

MSEIJ

View full text Add to dashboard Cite

Laser powder bed fusion method is popularly applied in the additive manufacturing of metal parts. The void defect and microstructure are the main factors which determine their mechanical properties. However, the characterization of microstructure and cavities is two dimensional, which is hard to show the spatial profile. In this paper, in order to explore the microstructure and defects in three dimensions, the combined continuously slicing and microstructure observation was used to investigate the microstructure of an Inconel 718 sample. The sample was sliced 468 layers with thickness of 1 µm by xenon ion beam, a 142.8 µm* 107 µm* 46.8 µm microstructure cube was reconstructed. From the 3D model, the melt pool, cavity, pore and grains and their orientations were analyzed. The results provide spatial features of its microstructure. The equi-axed grains are among the coarse column grains, and some are the original grains of insufficiently melt or totally unmelt powder particles. The results tell the difference of the two kinds of voids, i.e., cavity and pore.

show abstract

Section: Introductionmentioning

confidence: 93%

“…[13][14][15] The results show there is proper parameters or laser energy density. Zhang et al, 15 showed that the threshold laser energy density for the Al-Cu-Mg powder is 340 J/mm 3 . Too high energy density will also cause void defects as insufficient energy density.…”

Section: Introductionmentioning

confidence: 93%

3D characterization of the microstructure of LPBF- fabricated Inconel 718 alloy

Kang¹,

Huangb²,

Yub³

2023

MSEIJ

View full text Add to dashboard Cite

show abstract

“…In recent years, 3D printing has become a flexible and powerful technology in advanced manufacturing [ 1 ]. It has been widely used in many countries for personalized manufacturing and smart fabrication [ 2 , 3 ], which has also led to a huge change in industrial design [ 4 , 5 ], manufacturing methods [ 6 , 7 ], and digital flow [ 8 ]. With the diversification of materials and digitization of processes, 3D printing for high fidelity manufacturing of shape and performance is becoming more and more achievable [ 9 , 10 , 11 ].…”

Section: Introductionmentioning

confidence: 99%

Developing a Quality Evaluation System for Color Reproduction of Color 3D Printing Based on MATLAB Multi-Metrics

2023

View full text Add to dashboard Cite

Color 3D printing has been widely used in many fields such as cultural, medical, industrial, and food. The color reproduction accuracy of 3D printed products in these fields is becoming increasingly demanding, which requires more reproduction methods and practical tools. At present, most color 3D printing devices use one quantitative index, that is, color difference, to directly predict the color reproduction quality. However, this single quantitative index is not optimal for the curved surface of 3D printed color objects. Based on color evaluation principles, in this study, five new quantitative metrics consisting of color gamut comparison index, color SSIM index, color FSIM index, iCID index, and subjective scaling values are proposed for comparison, and the corresponding GUI design and code implementation of new color quality evaluation system are performed by MATLAB. Moreover, the comprehensive color assessment of color 3D printed products is confirmed by utilizing standard image acquisition and microscopic imaging methods that are not limited to printing materials and sampling locations. The operation of this system is validated to provide interactivity, simplicity and high efficiency. As a result, the system can provide new valuable feedback for color separation and output calibration of color 3D printing devices.

show abstract

“…However, the uneven distribution of laser absorption on the keyhole wall can lead to the formation of local hotspots, resulting in an imbalance between recoil pressure, vapor dynamics, capillary forces, and Marangoni forces [6]. As a result, the gas-liquid interface within the keyhole often exhibits strong fluctuations, which are affected by laser parameters (such as power, scanning velocity, spot size, duty cycle, etc.)…”

Section: Introductionmentioning

confidence: 99%

“…and the properties of the protective gas and substrate materials [7,8]. This significant variation in keyhole geometry can affect the energy coupling mechanism between the laser and the material, leading to unusual melt pool dynamics and the formation of spatters and pores [3,6,9,10]. Therefore, it is essential to estimate or measure the absorptance of laser light, as it is the energy source causing the keyhole formation.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning Approaches for Time-resolved Laser Absorptance Prediction in Additive Manufacturing

Jiang

Smith

et al. 2023

Preprint

View full text Add to dashboard Cite

The quantification of absorbed light is essential for understanding laser-material interactions and melt pool dynamics in order to minimize defects in additively manufactured metal components. The geometry of a vapor depression, also known as a keyhole, in melt pools formed during laser melting is closely related to laser energy absorption. This relationship has been observed by the state-of-the-art in situ high speed synchrotron x-ray visualization and integrating sphere radiometry. These two techniques create a temporally resolved dataset consisting of keyhole images and corresponding laser absorptance. In this work, we propose two different approaches to predict laser absorptance. The end-to-end approach uses deep convolutional neural networks to learn features of unprocessed x-ray images automatically without human supervision and predict the laser energy absorptance. The two-stage approach uses a semantic segmentation model to engineer geometric features and predict absorptance using classical regression models. Though with different advantages, both approaches reached a smooth mean absolute error less than 3.4%.

show abstract

Machine learning–aided real-time detection of keyhole pore generation in laser powder bed fusion

Cited by 74 publications

References 54 publications

3D characterization of the microstructure of LPBF- fabricated Inconel 718 alloy

3D characterization of the microstructure of LPBF- fabricated Inconel 718 alloy

Developing a Quality Evaluation System for Color Reproduction of Color 3D Printing Based on MATLAB Multi-Metrics

Deep Learning Approaches for Time-resolved Laser Absorptance Prediction in Additive Manufacturing

Contact Info

Product

Resources

About