Pith. sign in

REVIEW 44 references

Airborne acoustic emission enables sub-scanline keyhole porosity quantification and effective process characterization for metallic laser powder bed fusion

T0 review · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Airborne acoustic emission carries enough keyhole-dynamics information to quantify keyhole porosity at sub-scanline resolution, and the same signal can outline the keyhole regime on the power–velocity process map without X-ray CT.

desk verdict First credible AE regression to sub-scanline keyhole porosity density; labels are the soft spot, not the modeling. read the letter →

arxiv 2508.13492 v1 pith:CFBPQEUP submitted 2025-08-19 physics.app-ph

Haolin Liu , David Guirguis , Xuzhe Zeng , Logan Maurer , Vigknesh Rajan , Niloofar Sanaei , Chi-Ta Yang , Jack L. Beuth
show 2 more authors
Anthony D. Rollett Levent Burak Kara
This is my paper · ORCID
classification physics.app-ph
keywords laserpowderbedfusionkeyholeporosityacousticemissionKHLineNumX-raycomputedtomographyCNNregressionprocessmapTi-6Al-4V
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that airborne acoustic emission (AE) recorded during laser powder bed fusion contains enough keyhole-dynamics information to quantify keyhole-induced porosity at sub-scanline resolution. Its central move is a new metric, KHLineNum—the number of keyhole pores per unit scan length—which proves more learnable from AE scalograms than volumetric or total-count measures. A lightweight convolutional network fed with AE scalograms plus scan speed predicts KHLineNum with R² above 0.8 on single-bead validation, and the same model transfers to rasterized pad builds with an average R² of 0.924 for total pore count. The framework also derives isocontours of KHNum that match a previously reported keyhole-process-window boundary on the P–V map, suggesting a non-destructive, in situ route to process characterization. If correct, AE could replace much of the labor-intensive XCT workflow for keyhole porosity quantification and process-map building.

What carries the argument

The defining object is KHLineNum = N_pores / L_travel, a spatially resolved porosity metric expressed as pores per unit scan length, which the paper argues is the most discriminative target for AE-based regression. The carrying mechanism is a two-channel CNN input: a Morlet-wavelet continuous wavelet transform (CWT) scalogram of each AE snippet (5–50 kHz, log-frequency axis) plus the scan speed V tiled into a constant image, enabling the network to map spectrotemporal acoustic structure and process speed to local pore density.

What would settle it

Perform in-situ synchrotron X-ray imaging during the same SBD and PAD builds to record pore generation and elimination events in real time, then compare the imaged pore-creation locations against the XCT-derived KHLineNum labels used for training; a systematic mismatch—either more than 50 µm drift or remelting-induced pore disappearance within the snippet window—would show that the reported R² does not measure what the prose claims.

Watch

Extended reading notes

Core claim

The authors establish that KHLineNum—the number of keyhole pores per unit laser travel length—can be predicted from airborne AE scalograms combined with the scan speed V, using a five-layer CNN that takes a 224×224×2 input (the CWT scalogram image and a tiled, normalized V). On single-bead specimens the model achieves R² > 0.8 across snippet durations, with 6–8 ms windows giving the best accuracy–resolution trade-off; applying the SBD-trained model to pad specimens without fine-tuning yields an average R² of 0.924 for total KHNum. Ablation shows that masking the 37–45 kHz band degrades predictions most, aligning with known keyhole oscillation frequencies. Finally, level-set maps of predicted

Load-bearing premise

The regression labels assume that pores visible in the post-build X-ray scan are faithful, stationary records of pores generated while the laser passed that snippet's spatial window—within a fixed 50 µm drift tolerance—and that no pores are eliminated by remelting before imaging, an assumption the authors themselves flag as an unaccounted limitation in Section 5.

Editorial extensions

If this is right

  • Millisecond-scale keyhole porosity quantification becomes possible with a single airborne microphone, providing a basis for online monitoring of metal 3D printing.
  • The SBD-trained model generalizes to rasterized pad builds with different hatch spacings and scan orientations, suggesting geometry-agnostic AE-based defect regression.
  • The 37–45 kHz AE band is the most informative for keyhole pore prediction, connecting the data-driven result to the physical picture of keyhole oscillations.
  • AE-derived KHNum isocontours can approximate the keyhole process-window boundary on the P–V map, potentially replacing XCT-based process characterization for parameter selection.
  • A pre-trained model could in principle be applied to new alloys without XCT labeling, provided the keyhole dynamics are similar, enabling low-cost process-map reconstruction.
  • The framework could support fatigue-life assessment by linking predicted KH pore count to mechanical performance.
  • The spatial resolution of porosity localization (sub-scanline, milliseconds) opens a route toward closed-loop control that adjusts laser parameters when a pore-generation event is detected.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The reported R² values should be read as conditional on the labeling assumption that final XCT pore positions faithfully record pore generation events; if remelting eliminates or displaces pores beyond the 50 µm tolerance, the R² would overstate the true AE–porosity link—a testable concern the authors explicitly flag in Section 5.
  • The success of the two-channel input suggests that scan speed acts as a necessary context for interpreting AE, while including laser power P degraded performance (Appendix A); this hints that the AE scalogram already encodes the energy-density information P would add, an interaction worth probing with ablations that vary P while holding V fixed.
  • If the KHLineNum framework transfers across alloys, it could turn a single microphone into a rapid process-characterization tool for new materials, bypassing the XCT bottleneck entirely; the paper leaves this cross-material generalization unverified.
  • The sharp rise in KHNum between 600 and 1000 pores that marks the AE-derived KH-bound implies a detectable acoustic transition at the keyhole instability onset, which could be exploited as a real-time early-warning trigger during builds.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Circularity Check

1 steps flagged · score 3.0 of 10

Core R2 regression is not circular (XCT labels are external to AE), but the process-map 'AE-derived KH-bounds' claim is partially circular because the 600–1000 isocontour levels are selected from the same PAD ground-truth KHNum sharp rise.

  1. fitted input called prediction [Section 4, 'A potential application' (Fig. 17 discussion)]
    "We further examined the rationale behind selecting the KHNum=600 to KHNum=1000 isocontours as the AE-derived KH-bounds. ... To test this hypothesis, we visualized KHNum of the examined PAD builds in a 1-D plot. As shown in Fig. 17, a pronounced increase in KHNum is observed between 600 (red dashed line) and 1000 (blue dashed line), aligning closely with our chosen KHNum isocontours. This consistency confirms that the AE-derived KH-bounds capture the underlying physics of keyhole oscillation and instability, which ultimately governs the onset of KH porosity as well as the true KH-bound."

    The 'AE-derived KH-bounds' are not derived from AE alone: the isocontour levels (KHNum=600 to 1000) are chosen after inspecting the same PAD XCT ground-truth KHNum distribution (Fig. 17). The model's predicted isocontours at those hand-picked levels are then compared with Zhao's independently reported KH-bound and presented as matching. Because the threshold values were calibrated to the sharp rise in the very data used for validation, the agreement is partly by construction. This does not make the core R2 regression circular (labels come from XCT, not AE), but it overstates the claim of 'direct inference of KH regime boundaries' from AE: the level-set values were selected from the ground truth, not inferred from AE.

full rationale

The main supervised-learning result (Sec. 3.2, Fig. 8) is not circular: KHLineNum labels are computed from ex situ XCT pore counts via Eq. (1) (Npores/Ltravel), while the inputs are AE scalograms plus scan speed V. The labels are external to the AE modality, so the reported R2>0.8 is an empirical generalization claim, not a tautology. Although V appears both as an input and in the denominator of the label, V alone cannot determine Npores, and the paper's own ablation (Sec. 3.4.1) shows that V-only input degrades performance; thus the shared V does not force the prediction. The frequency-band analysis (Sec. 3.4.2) is a post-hoc interpretation with independent literature support, and the mask-one-band-out procedure is not circular. Self-citations (e.g., [6], [18], [25], [26]) are used for motivation or as external experimental references; Zhao et al. is an independent experimental process map despite overlapping authorship, so it is not load-bearing self-citation. The paper's own limitations about the fixed 50 µm drift tolerance and remelting-induced pore elimination (Secs. 2.2.2 and 5) are label-fidelity concerns that affect correctness, not circularity, because XCT remains an independent modality. The only notable circular step is in the process-map application: selecting the KHNum=600–1000 isocontours after observing the sharp rise in the same PAD ground-truth data, then calling them 'AE-derived KH-bounds,' makes the boundary-match claim partially a threshold-calibration artifact. Overall, the central claim retains independent content, so the circularity score is modest.

Assumptions & free parameters 5 free parameters · 6 assumptions · 1 invented entities

The ledger shows the central R² numbers rest on five hand-set or empirically calibrated constants: the drift tolerance, pore exclusion cutoffs, snippet duration floor, latency shift, and the post hoc isocontour level for the KH-bound. The regression itself is not circular (labels are external XCT), but every label passes through the drift and exclusion constants, and the boundary claim passes through the threshold selection. Domain assumptions about pore immobility, constant scan speed, transfer of the Zhao et al. boundary, and frequency-band interpretability are all reasonable but unverified within this paper. No hidden physical entities are postulated; KHLineNum is a definitional metric.

free parameters (5)
  • KH-bound isocontour level = KHNum = 600 to 1000 pores per PAD
    Sec. 4, Figs. 15-17. The isocontour levels presented as the AE-derived KH-bound were chosen by inspecting the sharp KHNum rise in the same PAD data, and the match with the Zhao et al. [26] boundary is then shown as validation. Post hoc threshold selection rather than an independently fixed level.
  • Pore drift tolerance = 50 um fixed, independent of scan speed
    Sec. 2.2.2 'KH pore drift accommodation'. Appended to every snippet's spatial window to relabel drifted pores. Chosen by hand because no in-situ synchrotron tracking was available; it directly controls which XCT pores label which AE snippet for every sample.
  • Pore exclusion thresholds = voxel count < 4 excluded; Feret diameter < 5.92 um excluded; 0.1 mm from scanline ends excluded
    Sec. 2.2.2. Empirically determined on the datasets used for training and validation to separate KH pores from segmentation artifacts and turnaround effects; applied uniformly, but the cutoffs influence every label.
  • Minimum snippet duration = 4 ms floor; 6-8 ms called optimal
    Sec. 2.2.2 and Fig. 8. Floor imposed to limit drift misattribution; the 4-10 ms blend used in training and the 'optimal' duration were selected on validation performance.
  • Acoustic latency shift = 72 samples (0.72 ms) uniformly applied
    Sec. 2.2.1. Uniform time shift of AE relative to photodiode. The stated 0.72 ms implies an acoustic path on the order of 0.25 m at argon sound speed, inconsistent with the stated 50 mm sensor distance, so the shift behaves as a tuned constant; any error misregisters pores to snippets.
assumptions (6)
  • domain assumption Final XCT pore positions are immobile records of generation events, plus a fixed 50 um drift allowance
    Sec. 2.2.2 registration: each pore is assigned to the AE snippet whose spatial window plus 50 um contains the pore's final position. Assumes no pore moves beyond 50 um and no pore is eliminated by remelting before scanning. The authors flag remelting-based elimination as unaccounted in Sec. 5.
  • domain assumption Nominal scan speed V is constant and accurate during steady portions of each scanline
    Sec. 2.2.2: L_travel = V * t converts snippet time to a spatial window; only turnaround regions (0.1 mm) are excluded, and no direct encoder verification of V is reported.
  • domain assumption Zhao et al. [26] KH-bound transfers to this machine (EOS M290) and material (Ti-64)
    Sec. 4 / Fig. 16: the reference boundary was measured with synchrotron X-ray imaging on different setups; its transfer to the M290 is assumed when judging the AE-derived isocontours.
  • domain assumption Keyhole oscillation frequencies from prior work [11, 39] match this machine's AE band
    Sec. 3.4.2 and Sec. 5: the 37-45 kHz informative band is interpreted through Tempelman et al. [11] and Khairallah et al. [39]; transferability of those frequencies to this sensor, machine, and geometry is assumed.
  • domain assumption Dragonfly segmentation thresholds correctly separate KH pores from artifacts
    Sec. 2.2.2: global intensity thresholding, connected components, morphological filtering, and manual verification; no quantitative segmentation accuracy check against a reference standard is reported.
  • standard math Morlet CWT scalogram is an adequate time-frequency representation of the AE snippets
    Sec. 2.2.1 uses ssqueezepy with a Morlet mother wavelet; standard signal-processing method, but no formal justification of its adequacy for this signal class is given.
invented entities (1)
  • KHLineNum
    purpose: Regression target: number of KH pores per unit scan length (Eq. 1)
    A definitional metric, not a hidden physical entity; its values are inherited entirely from the XCT segmentation and the drift and exclusion constants, so it adds no independent falsifiable handle beyond the data pipeline it is built from.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Airborne acoustic emission enables sub-scanline keyhole porosity quantification and effective process characterization for metallic laser powder bed fusion." pith.science (2026). https://pith.science/paper/CFBPQEUP

@misc{pith2026250813492,
  author       = {Pith},
  title        = {Pith review of: Airborne acoustic emission enables sub-scanline keyhole porosity quantification and effective process characterization for metallic laser powder bed fusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CFBPQEUP}},
  note         = {Machine review of arXiv:2508.13492}
}
read the original abstract

Keyhole-induced (KH) porosity, which arises from unstable vapor cavity dynamics under excessive laser energy input, remains a significant challenge in laser powder bed fusion (LPBF). This study presents an integrated experimental and data-driven framework using airborne acoustic emission (AE) to achieve high-resolution quantification of KH porosity. Experiments conducted on an LPBF system involved in situ acquisition of airborne AE and ex situ porosity imaging via X-ray computed tomography (XCT), synchronized spatiotemporally through photodiode signals with submillisecond precision. We introduce KHLineNum, a spatially resolved porosity metric defined as the number of KH pores per unit scan length, which serves as a physically meaningful indicator of the severity of KH porosity in geometries and scanning strategies. Using AE scalogram data and scan speed, we trained a lightweight convolutional neural network to predict KHLineNum with millisecond-scale temporal resolution, achieving an R-squared value exceeding 0.8. Subsequent analysis identified the 35-45 kHz frequency band of AE as particularly informative, consistent with known KH oscillations. Beyond defect quantification, the framework also enables AE-driven direct inference of KH regime boundaries on the power-velocity process map, offering a noninvasive and scalable component to labor-intensive post-process techniques such as XCT. We believe this framework advances AE-based monitoring in LPBF, providing a pathway toward improved quantifiable defect detection and process control.

Figures

Figures reproduced from arXiv: 2508.13492 by the authors.

Figure 1
Figure 1. Rationale and motivation for acoustic-based KH porosity quantification. The green line represents one of [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. An overview of experimental details. (a) LPBF machine and the associated sensor setups. (b) Details of the [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Red points represent the selected P−V points used for the LPBF bare-plate experiments. The blue dashed line is a nominal keyhole-process-window boundary reported by Zhao et al. [26]. tion and analysis were performed in Dragonfly [27], yielding detailed porosity characteristics— including size, location, morphology, etc.—which were exported for subsequent training and eval￾uation of the ML model. (see [PITH_FULL_IMA… view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: Temporal clipping of an AE snippet and the corresponding AE scalogram snippet after CWT processing. [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Spatiotemporal registration between AE scalogram snippets and the corresponding XCT porosity image [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: A tractable ML architecture for AE-driven [PITH_FULL_IMAGE:figures/full_fig_p017_6.png]
Figure 7
Figure 7. Figure 7: Qualitative correlation analysis between [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: R 2 plots of KHLineNum predictions across varying AE snippet durations for SBDs. A gradual decline in prediction accuracy is observed as snippet duration decreases. to three factors: (1) the model’s tendency to misinterpret random noise as signal, (2) the transient for…
Figure 9
Figure 9. Figure 9: Temporal tracking of KH porosity generation for an SBD from the test dataset using a non-overlapped 12 ms [PITH_FULL_IMAGE:figures/full_fig_p021_9.png]
Figure 10
Figure 10. Figure 10: Violin plots comparing the log-transformed prediction error (log [PITH_FULL_IMAGE:figures/full_fig_p022_10.png]
Figure 11
Figure 11. Figure 11: R 2 plots of KHNum predictions for PAD dataset with different time window durations. Colors and marker types represent P and V of the fusion laser, respectively. As shown in [PITH_FULL_IMAGE:figures/full_fig_p023_11.png]
Figure 12
Figure 12. Figure 12: Investigation strategy for the ablation studies conducted by altering the model input. The red pathway [PITH_FULL_IMAGE:figures/full_fig_p024_12.png]
Figure 13
Figure 13. Figure 13: The prediction performance of both ablated models on the SBD test dataset (top row) as well as the entire [PITH_FULL_IMAGE:figures/full_fig_p025_13.png]
Figure 14
Figure 14. Figure 14: RMSE-based feature importance analysis using the fixed pre-trained baseline model on various MOBO [PITH_FULL_IMAGE:figures/full_fig_p026_14.png]
Figure 15
Figure 15. Figure 15: Heat map and isocontours of KHNum generated using the curated PAD dataset and the pre-trained baseline model. Note that the shape of the isocontours may vary depending on the distribution of selected P−V points. We aim to illustrate that, by leveraging the AE scalogra…
Figure 16
Figure 16. Figure 16: AE-derived KH-bounds (the selected red KHNum isocontours from [PITH_FULL_IMAGE:figures/full_fig_p029_16.png]
Figure 17
Figure 17. Figure 17: 1-D plot of KHNum for all PADs. The gap between 600 (red dashed line) and 1000 (blue dashed line) matches our selection for the AE-derived KH-bounds. If further validated across materials, this methodology may offer a foundation for generalizing AE-based KH porosity i…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 38 canonical work pages

  1. [1]

    J. V . Gordon, S. P. Narra, R. W. Cunningham, H. Liu, H. Chen, R. M. Suter, J. L. Beuth, A. D. Rollett, Defect structure process maps for laser powder bed fusion additive manufacturing, Additive Manufacturing 36 (2020) 101552

    work page 2020

  2. [2]

    W. H. Kan, L. N. S. Chiu, C. V . S. Lim, Y . Zhu, Y . Tian, D. Jiang, A. Huang, A critical review 32 on the effects of process-induced porosity on the mechanical properties of alloys fabricated by laser powder bed fusion, Journal of Materials Science 57 (2022) 9818–9865

    work page 2022

  3. [3]

    L. Guo, H. Liu, H. Wang, Q. Wei, Y . Xiao, Z. Tang, Y . Wu, H. Wang, Identifying the keyhole stability and pore formation mechanisms in laser powder bed fusion additive manufacturing, Journal of Materials Processing Technology 321 (2023) 118153

    work page 2023

  4. [4]

    T. M. Mower, M. J. Long, Mechanical behavior of additive manufactured, powder-bed laser- fused materials, Materials Science and Engineering: A 651 (2016) 198–213

    work page 2016

  5. [5]

    Edwards, M

    P. Edwards, M. Ramulu, Fatigue performance evaluation of selective laser melted ti–6al–4v, Materials Science and Engineering: A 598 (2014) 327–337

    work page 2014

  6. [6]

    Z. Ren, J. Shao, H. Liu, S. J. Clark, L. Gao, L. Balderson, K. Mumm, K. Fezzaa, A. D. Rollett, L. B. Kara, et al., Sub-millisecond keyhole pore detection in laser powder bed fusion using sound and light sensors and machine learning, Materials Futures 3 (2024) 045001

    work page 2024

  7. [7]

    Du Plessis, Effects of process parameters on porosity in laser powder bed fusion revealed by x-ray tomography, Additive Manufacturing 30 (2019) 100871

    A. Du Plessis, Effects of process parameters on porosity in laser powder bed fusion revealed by x-ray tomography, Additive Manufacturing 30 (2019) 100871

    work page 2019

  8. [8]

    Forien, N

    J.-B. Forien, N. P. Calta, P. J. DePond, G. M. Guss, T. T. Roehling, M. J. Matthews, Detecting keyhole pore defects and monitoring process signatures during laser powder bed fusion: A correlation between in situ pyrometry and ex situ x-ray radiography, Additive Manufacturing 35 (2020) 101336

    work page 2020

Show all 44 references
  1. [9]

    Wahlquist, A

    S. Wahlquist, A. Ali, Roles of modeling and artificial intelligence in lpbf metal print defect detection: Critical review, Applied Sciences 14 (2024) 8534

    2024

  2. [10]

    L. Chen, X. Yao, C. Tan, W. He, J. Su, F. Weng, Y . Chew, N. P. H. Ng, S. K. Moon, In-situ crack and keyhole pore detection in laser directed energy deposition through acoustic signal and deep learning, Additive Manufacturing 69 (2023) 103547. 33

    2023

  3. [11]

    J. R. Tempelman, A. J. Wachtor, E. B. Flynn, P. J. Depond, J.-B. Forien, G. M. Guss, N. P. Calta, M. J. Matthews, Detection of keyhole pore formations in laser powder-bed fusion using acoustic process monitoring measurements, Additive Manufacturing 55 (2022) 102735

    2022

  4. [12]

    J. R. Tempelman, A. J. Wachtor, E. B. Flynn, P. J. Depond, J.-B. Forien, G. M. Guss, N. P. Calta, M. J. Matthews, Sensor fusion of pyrometry and acoustic measurements for local- ized keyhole pore identification in laser powder bed fusion, Journal of Materials Processing Techno...

    2022

  5. [13]

    Guillen, S

    D. Guillen, S. Wahlquist, A. Ali, Critical review of lpbf metal print defects detection: roles of selective sensing technology, Applied Sciences 14 (2024) 6718

    2024

  6. [14]

    J. R. Tempelman, M. K. Mudunuru, S. Karra, A. J. Wachtor, B. Ahmmed, E. B. Flynn, J.-B. Forien, G. M. Guss, N. P. Calta, P. J. DePond, et al., Uncovering acoustic signatures of pore formation in laser powder bed fusion, The International Journal of Advanced Manufacturing Techn...

    2024

  7. [15]

    Drissi-Daoudi, V

    R. Drissi-Daoudi, V . Pandiyan, R. Logé, S. Shevchik, G. Masinelli, H. Ghasemi-Tabasi, A. Parrilli, K. Wasmer, Differentiation of materials and laser powder bed fusion process- ing regimes from airborne acoustic emission combined with machine learning, Virtual and Physical Pro...

    2022

  8. [16]

    S. A. Shevchik, C. Kenel, C. Leinenbach, K. Wasmer, Acoustic emission for in situ quality monitoring in additive manufacturing using spectral convolutional neural networks, Additive Manufacturing 21 (2018) 598–604

    2018

  9. [17]

    S. A. Shevchik, G. Masinelli, C. Kenel, C. Leinenbach, K. Wasmer, Deep learning for in situ and real-time quality monitoring in additive manufacturing using acoustic emission, IEEE Transactions on Industrial Informatics 15 (2019) 5194–5203

    2019

  10. [18]

    H. Liu, C. Gobert, K. Ferguson, B. Abranovic, H. Chen, J. L. Beuth, A. D. Rollett, L. B. Kara, Inference of highly time-resolved melt pool visual characteristics and spatially-dependent 34 lack-of-fusion defects in laser powder bed fusion using acoustic and thermal emission da...

    2024

  11. [19]

    Hamidi Nasab, G

    M. Hamidi Nasab, G. Masinelli, C. de Formanoir, L. Schlenger, S. Van Petegem, R. Es- maeilzadeh, K. Wasmer, A. Ganvir, A. Salminen, F. Aymanns, et al., Harmonizing sound and light: X-ray imaging unveils acoustic signatures of stochastic inter-regime instabilities during laser ...

    2023

  12. [20]

    Wasmer, C

    K. Wasmer, C. Kenel, C. Leinenbach, S. Shevchik, In situ and real-time monitoring of powder-bed am by combining acoustic emission and artificial intelligence, in: Indus- trializing Additive Manufacturing-Proceedings of Additive Manufacturing in Products and Applications-AMPA20...

    2018

  13. [21]

    Pandiyan, R

    V . Pandiyan, R. Drissi-Daoudi, S. Shevchik, G. Masinelli, T. Le-Quang, R. Loge, K. Wasmer, Deep transfer learning of additive manufacturing mechanisms across materials in metal-based laser powder bed fusion process, Journal of Materials Processing Technology 303 (2022) 117531

    2022

  14. [22]

    Drissi Daoudi, Towards Robust Monitoring of the Laser Powder Bed Fusion Process based on Acoustic Emission combined with Machine Learning Solutions, Ph.D

    R. Drissi Daoudi, Towards Robust Monitoring of the Laser Powder Bed Fusion Process based on Acoustic Emission combined with Machine Learning Solutions, Ph.D. thesis, EPFL, 2023

    2023

  15. [23]

    R. Liu, S. Liu, X. Zhang, A physics-informed machine learning model for porosity analysis in laser powder bed fusion additive manufacturing, The International Journal of Advanced Manufacturing Technology 113 (2021) 1943–1958

    2021

  16. [24]

    URL:https: //www.cmu.edu/me/xctf/xrayct/index.html, accessed: 2025-07-25

    Carnegie Mellon University, X-ray computed tomography facility (xctf), n.d. URL:https: //www.cmu.edu/me/xctf/xrayct/index.html, accessed: 2025-07-25

    2025

  17. [25]

    Z. Ren, L. Gao, S. J. Clark, K. Fezzaa, P. Shevchenko, A. Choi, W. Everhart, A. D. Rollett, L. Chen, T. Sun, Machine learning–aided real-time detection of keyhole pore generation in laser powder bed fusion, Science 379 (2023) 89–94. 35

    2023

  18. [26]

    C. Zhao, N. D. Parab, X. Li, K. Fezzaa, W. Tan, A. D. Rollett, T. Sun, Critical instability at moving keyhole tip generates porosity in laser melting, Science 370 (2020) 1080–1086

    2020

  19. [27]

    Available fromhttps://www.theobjects

    Object Research Systems (ORS) Inc., Dragonfly (version 2024.1) [software],https: //www.theobjects.com/dragonfly, 2024. Available fromhttps://www.theobjects. com/dragonfly

    2024

  20. [28]

    Z. Qin, P. Zhang, F. Wu, X. Li, Fcanet: Frequency channel attention networks, in: Proceed- ings of the IEEE/CVF international conference on computer vision, 2021, pp. 783–792

    2021

  21. [29]

    Muradeli, ssqueezepy, GitHub

    J. Muradeli, ssqueezepy, GitHub. Note: https://github.com/OverLordGoldDragon/ssqueezepy/ (2020). doi:10.5281/zenodo.5080508

    2020 doi

  22. [30]

    S. Pang, W. Chen, W. Wang, A quantitative model of keyhole instability induced porosity in laser welding of titanium alloy, Metallurgical and Materials Transactions A 45 (2014) 2808–2818

    2014

  23. [31]

    Van Rossum, F

    G. Van Rossum, F. L. Drake Jr, Python reference manual, Centrum voor Wiskunde en Infor- matica Amsterdam, 1995

    1995

  24. [32]

    C. R. Harris, K. J. Millman, S. J. van der Walt, R. Gommers, P. Virtanen, D. Courna- peau, E. Wieser, J. Taylor, S. Berg, N. J. Smith, R. Kern, M. Picus, S. Hoyer, M. H. van Kerkwijk, M. Brett, A. Haldane, J. F. del Río, M. Wiebe, P. Peterson, P. Gérard-Marchant, K. Sheppard, ...

    2020 doi

  25. [33]

    Virtanen, R

    P. Virtanen, R. Gommers, T. E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, E. Burovski, P. Peterson, W. Weckesser, J. Bright, S. J. van der Walt, M. Brett, J. Wilson, K. J. Millman, N. Mayorov, A. R. J. Nelson, E. Jones, R. Kern, E. Larson, C. J. Carey, ˙I. Po- lat, Y . F...

    2020

  26. [34]

    Wes McKinney, Data Structures for Statistical Computing in Python, in: Stéfan van der Walt, Jarrod Millman (Eds.), Proceedings of the 9th Python in Science Conference, 2010, pp. 56 – 61. doi:10.25080/Majora-92bf1922-00a

    2010 doi

  27. [35]

    J. D. Hunter, Matplotlib: A 2d graphics environment, Computing in Science & Engineering 9 (2007) 90–95. doi:10.1109/MCSE.2007.55

    2007 doi

  28. [36]

    Clark, Pillow (pil fork) documentation, 2015

    A. Clark, Pillow (pil fork) documentation, 2015

    2015

  29. [37]

    Paszke, S

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, S. Chintala, Pytorch: An imperative style, high- perfor...

    2019

  30. [38]

    W. Wang, J. Ning, S. Y . Liang, Analytical prediction of keyhole porosity in laser powder bed fusion, The International Journal of Advanced Manufacturing Technology 119 (2022) 6995–7002

    2022

  31. [39]

    S. A. Khairallah, T. Sun, B. J. Simonds, Onset of periodic oscillations as a precursor of a transition to pore-generating turbulence in laser melting, Additive Manufacturing Letters 1 (2021) 100002

    2021

  32. [40]

    Huang, T

    Y . Huang, T. G. Fleming, S. J. Clark, S. Marussi, K. Fezzaa, J. Thiyagalingam, C. L. A. Leung, P. D. Lee, Keyhole fluctuation and pore formation mechanisms during laser powder bed fusion additive manufacturing, Nature communications 13 (2022) 1170. 37

    2022

  33. [41]

    de Formanoir, M

    C. de Formanoir, M. H. Nasab, L. Schlenger, S. Van Petegem, G. Masinelli, F. Marone, A. Salminen, A. Ganvir, K. Wasmer, R. E. Loge, Healing of keyhole porosity by means of defocused laser beam remelting: Operando observation by x-ray imaging and acoustic emission-based detecti...

    2024

  34. [42]

    Reddy, A

    T. Reddy, A. Ngo, J. P. Miner, C. Gobert, J. L. Beuth, A. D. Rollett, J. J. Lewandowski, S. P. Narra, Fatigue-based process window for laser beam powder bed fusion additive manufac- turing, International Journal of Fatigue 187 (2024) 108428

    2024

  35. [43]

    K. He, X. Zhang, S. Ren, J. Sun, Deep residual learning for image recognition, in: Proceed- ings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778

    2016

  36. [44]

    Dosovitskiy, L

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. De- hghani, M. Minderer, G. Heigold, S. Gelly, et al., An image is worth 16x16 words: Trans- formers for image recognition at scale, arXiv preprint arXiv:2010.11929 (2020). 38 Appendix A. Addi...

    2010 arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.