pith. machine review for the scientific record. sign in

arxiv: 2604.09889 · v1 · submitted 2026-04-10 · 💻 cs.AI

Recognition: unknown

In-situ process monitoring for defect detection in wire-arc additive manufacturing: an agentic AI approach

Pallock Halder , Satyajit Mojumder

Authors on Pith no claims yet

Pith reviewed 2026-05-10 16:47 UTC · model grok-4.3

classification 💻 cs.AI
keywords additive manufacturingdefect detectionmulti-agent AIin-situ monitoringporosityacoustic signalsprocess monitoringwire arc additive manufacturing
0
0 comments X

The pith

A multi-agent AI framework using welder signals and acoustic data detects porosity defects in wire-arc additive manufacturing with 91.6 percent accuracy.

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

This paper develops an agentic AI approach for monitoring defects during wire-arc additive manufacturing. Separate agents analyze welder process signals such as current and voltage, and acoustic emissions, each trained against X-ray computed tomography data to identify porosity. These agents are then coordinated by a large language model in a multi-agent setup that makes joint decisions. The combined system achieves better results than any agent working alone, with 91.6 percent decision accuracy and an F1 score of 0.821 across multiple runs. This could support real-time adjustments to produce better quality parts without waiting for post-process inspection.

Core claim

The paper claims that an LLM-orchestrated multi-agent system, consisting of a processing agent from welder current and voltage signals and a monitoring agent from acoustic data, both classified against XCT ground truth for porosity, outperforms individual agents by reaching 91.6% decision accuracy and 0.821 F1 score on decided runs over 15 independent trials, with average reasoning quality of 3.74 out of 5.

What carries the argument

The LLM-based orchestration of parallel processing and monitoring agents that fuses decisions from welder signals and acoustic data for in-situ porosity defect classification.

If this is right

  • The multi-agent configuration delivers higher accuracy and F1 scores than any single-agent setup.
  • Evaluation metrics show consistent performance across 15 independent runs for the coordinated system.
  • Both signal-based and acoustic agents can be trained effectively using XCT as ground truth for defect labeling.
  • The framework supports autonomous real-time monitoring toward qualified part production in WAAM and similar processes.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Deploying this in production lines might allow immediate correction of process parameters to avoid defects.
  • Extending the agents to predict other defect types like cracks or lack of fusion could broaden its utility.
  • Replacing the LLM orchestrator with a lighter model might enable faster on-machine deployment.

Load-bearing premise

Welder process signals and acoustic data hold sufficient information for reliable porosity classification based on XCT labels, and the LLM orchestration remains consistent beyond the current experimental dataset.

What would settle it

Apply the trained multi-agent system to a fresh set of WAAM builds with known porosity levels from XCT and measure if the accuracy stays above 85 percent or drops significantly.

Figures

Figures reproduced from arXiv: 2604.09889 by Pallock Halder, Satyajit Mojumder.

Figure 1
Figure 1. Figure 1: System overview of the proposed agentic AI framework for in-situ WAAM process monitoring. The architecture comprises three functional layers: (a) a signal acquisition layer capturing processing signals (welding current and voltage) and monitoring signals (acoustic data), (b) a machine learning tool layer with a 1D-CNN for processing signal classification and a 2D-CNN for monitoring signal (audio spectrogra… view at source ↗
Figure 2
Figure 2. Figure 2: LangGraph workflow graphs for the four agentic orchestration configurations. (a) Processing￾agent architecture: a single processing agent invokes the signal tool and performs reasoning on process signals, followed by extraction, reflection-based review, and report generation. (b) Monitoring-agent architecture: a single monitoring agent invokes the audio tool to analyze acoustic data, followed by extraction… view at source ↗
Figure 4
Figure 4. Figure 4: Class-wise monitoring score (MS) across different agent configurations. MS reflects the practical usefulness of monitoring recommendations. 3.3 Monitoring utility The practical value of a process monitoring system depends on two complementary properties: its ability to flag every defective track for human review, and its ability to leave normal tracks undisturbed. A system that indiscriminately flags all i… view at source ↗
read the original abstract

AI agents are being increasingly deployed across a wide range of real-world applications. In this paper, we propose an agentic AI framework for in-situ process monitoring for defect detection in wire-arc additive manufacturing (WAAM). The autonomous agent leverages a WAAM process monitoring dataset and a trained classification tool to build AI agents and uses a large language model (LLM) for in-situ process monitoring decision-making for defect detection. A processing agent is developed based on welder process signals, such as current and voltage, and a monitoring agent is developed based on acoustic data collected during the process. Both agents are tasked with identifying porosity defects from processing and monitoring signals, respectively. Ground truth X-ray computed tomography (XCT) data are used to develop classification tools for both the processing and monitoring agents. Furthermore, a multi-agent framework is demonstrated in which the processing and monitoring agents are orchestrated together for parallel decision-making on the given task of defect classification. Evaluation metrics are proposed to determine the efficacy of both individual agents, the combined single-agent, and the coordinated multi-agent system. The multi-agent configuration outperforms all individual-agent counterparts, achieving a decision accuracy of 91.6% and an F1 score of 0.821 on decided runs, across 15 independent runs, and a reasoning quality score of 3.74 out of 5. These in-situ process monitoring agents hold significant potential for autonomous real-time process monitoring and control toward building qualified parts for WAAM and other additive manufacturing processes.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

4 major / 2 minor

Summary. The manuscript proposes an agentic AI framework for in-situ process monitoring and defect detection in wire-arc additive manufacturing (WAAM). It develops a processing agent from welder current/voltage signals and a monitoring agent from acoustic data, each using classification tools trained against XCT ground truth for porosity identification. An LLM-orchestrated multi-agent system is introduced for parallel decision-making; this configuration is reported to outperform single-agent baselines, reaching 91.6% decision accuracy and 0.821 F1 on decided runs across 15 independent runs together with a reasoning quality score of 3.74/5.

Significance. If the empirical results are reproducible, the work offers a concrete demonstration of multi-agent orchestration for combining complementary in-situ signals in additive manufacturing. The use of independent XCT labels for both training and evaluation is a clear strength that avoids circularity. The reported lift from multi-agent coordination and the explicit reasoning-quality metric are also positive features. However, the overall significance is limited by the absence of core methodological details needed to judge whether the signals truly support reliable XCT-aligned classification or whether the gains generalize beyond the specific dataset and prompt set.

major comments (4)
  1. The methods description provides no information on dataset size, number of WAAM builds, train/test split ratios, or cross-validation procedure used to develop the classification tools for the processing and monitoring agents. These details are load-bearing for the central claim that the multi-agent system achieves 91.6% accuracy and 0.821 F1, because without them it is impossible to rule out overfitting or label noise as the source of the reported performance.
  2. No description is given of the feature engineering steps or the specific model architecture/type (e.g., random forest, neural net, or other) employed by the processing agent on current/voltage time series and by the monitoring agent on acoustic signals. This omission directly affects the weakest assumption that the in-situ signals contain sufficient information to classify XCT-labeled porosity.
  3. The temporal alignment procedure between the in-situ process signals (current, voltage, acoustic) and the XCT-detected pores is not specified. Without an explicit alignment method, the ground-truth mapping used to train and evaluate both agents cannot be verified, undermining the validity of the multi-agent decision accuracy numbers.
  4. The criteria used to designate “decided runs” in the multi-agent orchestration and the exact protocol for computing the reasoning quality score (3.74/5) are not stated, nor is any test of stability under prompt variation or across independent builds. These elements are central to the claim that the LLM layer adds reliable value beyond the individual agents.
minor comments (2)
  1. The abstract and evaluation section would benefit from a concise table summarizing the 15 runs (e.g., number of samples per run, decided vs. undecided counts) to make the performance numbers immediately interpretable.
  2. Notation for the two agents is occasionally inconsistent (processing agent vs. welder agent); a single, clearly defined terminology would improve readability.

Simulated Author's Rebuttal

4 responses · 0 unresolved

We thank the referee for their thorough review and constructive comments. We agree that several methodological details were insufficiently described in the original submission and will revise the manuscript to address these points. Our responses to each major comment are provided below.

read point-by-point responses

  1. Referee: The methods description provides no information on dataset size, number of WAAM builds, train/test split ratios, or cross-validation procedure used to develop the classification tools for the processing and monitoring agents. These details are load-bearing for the central claim that the multi-agent system achieves 91.6% accuracy and 0.821 F1, because without them it is impossible to rule out overfitting or label noise as the source of the reported performance.

    Authors: We acknowledge this omission. In the revised manuscript, we will add a comprehensive description of the dataset, including the total number of samples collected from multiple WAAM builds, the train/test split ratios employed, and the cross-validation strategy used to train and validate the classification tools for both agents. This will enable readers to better evaluate the reliability of the reported performance metrics. revision: yes

  2. Referee: No description is given of the feature engineering steps or the specific model architecture/type (e.g., random forest, neural net, or other) employed by the processing agent on current/voltage time series and by the monitoring agent on acoustic signals. This omission directly affects the weakest assumption that the in-situ signals contain sufficient information to classify XCT-labeled porosity.

    Authors: We agree that these details are important. We will expand the Methods section to detail the feature engineering process for both the current/voltage signals and the acoustic signals, as well as the specific machine learning models and architectures used for each agent. This will clarify how the in-situ signals are processed to align with the XCT ground truth. revision: yes

  3. Referee: The temporal alignment procedure between the in-situ process signals (current, voltage, acoustic) and the XCT-detected pores is not specified. Without an explicit alignment method, the ground-truth mapping used to train and evaluate both agents cannot be verified, undermining the validity of the multi-agent decision accuracy numbers.

    Authors: We will include a detailed explanation of the temporal alignment procedure in the revised paper. This will describe how the process signals are synchronized with the XCT data to establish the ground truth labels for porosity defects. revision: yes

  4. Referee: The criteria used to designate “decided runs” in the multi-agent orchestration and the exact protocol for computing the reasoning quality score (3.74/5) are not stated, nor is any test of stability under prompt variation or across independent builds. These elements are central to the claim that the LLM layer adds reliable value beyond the individual agents.

    Authors: We will clarify the criteria for designating 'decided runs' (such as agreement between agents or confidence thresholds) and provide the exact protocol for the reasoning quality score, which is based on expert evaluation. Additionally, we will include results on stability under prompt variations and across the 15 independent runs to support the robustness of the multi-agent approach. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical metrics rely on independent XCT ground truth

full rationale

The paper describes an empirical agentic framework where processing and monitoring agents are trained on in-situ signals (current/voltage, acoustics) with classification tools developed from separate XCT ground-truth labels. Reported performance (91.6% accuracy, 0.821 F1 on decided runs across 15 runs) is measured directly against these external labels rather than derived by construction from fitted parameters, self-definitions, or self-citation chains. No equations appear, no uniqueness theorems are invoked, and the multi-agent orchestration lift is presented as an observed outcome, not a mathematical reduction to inputs. The evaluation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The central claim depends on the existence of trained classification tools that map signals to XCT labels and on the LLM's ability to integrate agent outputs; no explicit free parameters, axioms, or new physical entities are stated in the abstract.

pith-pipeline@v0.9.0 · 5575 in / 1168 out tokens · 59534 ms · 2026-05-10T16:47:28.975392+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

115 extracted references · 104 canonical work pages · 13 internal anchors

  1. [1]

    Results This section presents the quantitative and qualitative performance of individual processing and monitoring agents, a combined single -agent and multi-agent system for i n-situ monitoring of WAAM. Performance is evaluated across classification tool performance, decision accuracy, monitoring utility, reasoning quality, and agentic behavior, includin...

  2. [2]

    This difference likely reflects what each sensor physically measures

    Discussion 4.1 Relative performance of signal modalities The monitoring agent substantially outperformed the processing agent under the conditions of this study, suggesting that a udio signals carry richer porosity- relevant information than processing signals. This difference likely reflects what each sensor physically measures . Monitoring signals captu...

  3. [3]

    Conclusion This paper presented an agentic AI framework for in- situ process monitoring of WAAM, integrating a 1D-CNN for processing signals and a 2D-CNN for monitoring signals classification as callable tools within a LangGraph -based multi-agent system powered by an LLM brain. Four configurations were evaluated: a processing agent, a monitoring agent, a...

  4. [4]

    D.D. Gu, W. Meiners, K. Wissenbach, R. Poprawe, Laser additive manufacturing of metallic components: materials, processes and mechanisms, International Materials Reviews 57 (2012) 133–164. https://doi.org/10.1179/1743280411Y.0000000014

    work page doi:10.1179/1743280411y.0000000014 2012

  5. [5]

    DebRoy, H.L

    T. DebRoy, H.L. Wei, J.S. Zuback, T. Mukherjee, J.W. Elmer, J.O. Milewski, A.M. Beese, A. Wilson-Heid, A. De, W. Zhang, Additive manufacturing of metallic components – Process, structure and properties, Progress in Materials Science 92 (2018) 112–224. https://doi.org/10.1016/j.pmatsci.2017.10.001

    work page doi:10.1016/j.pmatsci.2017.10.001 2018

  6. [6]

    Williams, F

    S.W. Williams, F. Martina, A.C. Addison, J. Ding, G. Pardal, P. Colegrove, Wire + Arc Additive Manufacturing, Materials Science and Technology 32 (2016) 641–647. https://doi.org/10.1179/1743284715Y.0000000073

    work page doi:10.1179/1743284715y.0000000073 2016

  7. [7]

    Rodrigues, V

    T.A. Rodrigues, V. Duarte, R.M. Miranda, T.G. Santos, J.P. Oliveira, Current Status and Perspectives on Wire and Arc Additive Manufacturing (WAAM), Materials 12 (2019) 1121. https://doi.org/10.3390/ma12071121

    work page doi:10.3390/ma12071121 2019

  8. [8]

    D. Ding, Z. Pan, D. Cuiuri, H. Li, Wire-feed additive manufacturing of metal components: technologies, developments and future interests, Int J Adv Manuf Technol 81 (2015) 465–481. https://doi.org/10.1007/s00170-015-7077-3

    work page doi:10.1007/s00170-015-7077-3 2015

  9. [9]

    B. Wu, Z. Pan, D. Ding, D. Cuiuri, H. Li, J. Xu, J. Norrish, A review of the wire arc additive manufacturing of metals: properties, defects and quality improvement, Journal of Manufacturing Processes 35 (2018) 127–139. https://doi.org/10.1016/j.jmapro.2018.08.001

    work page doi:10.1016/j.jmapro.2018.08.001 2018

  10. [10]

    Spencer, P.M

    J.D. Spencer, P.M. Dickens, C.M. Wykes, Rapid prototyping of metal parts by three-dimensional welding, Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture 212 (1998) 175–182. https://doi.org/10.1243/0954405981515590

    work page doi:10.1243/0954405981515590 1998

  11. [11]

    Martina, J

    F. Martina, J. Mehnen, S.W. Williams, P. Colegrove, F. Wang, Investigation of the benefits of plasma deposition for the additive layer manufacture of Ti–6Al–4V, Journal of Materials Processing Technology 212 (2012) 1377–1386. https://doi.org/10.1016/j.jmatprotec.2012.02.002

    work page doi:10.1016/j.jmatprotec.2012.02.002 2012

  12. [12]

    Prado-Cerqueira, J.L

    J.L. Prado-Cerqueira, J.L. Diéguez, A.M. Camacho, Preliminary development of a Wire and Arc Additive Manufacturing system (WAAM), Procedia Manufacturing 13 (2017) 895–902. https://doi.org/10.1016/j.promfg.2017.09.154

    work page doi:10.1016/j.promfg.2017.09.154 2017

  13. [13]

    Y. Li, C. Su, J. Zhu, Comprehensive review of wire arc additive manufacturing: Hardware system, physical process, monitoring, property characterization, application and future prospects, Results in Engineering 13 (2022) 100330. https://doi.org/10.1016/j.rineng.2021.100330

    work page doi:10.1016/j.rineng.2021.100330 2022

  14. [14]

    Zhang, R

    H. Zhang, R. Li, J. Liu, K. Wang, Q. Weijian, L. Shi, L. Lei, W. He, S. Wu, State-of-art review on the process-structure-properties-performance linkage in wire arc additive manufacturing, Virtual and Physical Prototyping 19 (2024) e2390495. https://doi.org/10.1080/17452759.2024.2390495

    work page doi:10.1080/17452759.2024.2390495 2024

  15. [15]

    Tomar, S

    B. Tomar, S. Shiva, T. Nath, A review on wire arc additive manufacturing: Processing parameters, defects, quality improvement and recent advances, Materials Today Communications 31 (2022) 103739. https://doi.org/10.1016/j.mtcomm.2022.103739

    work page doi:10.1016/j.mtcomm.2022.103739 2022

  16. [16]

    Colegrove, H.E

    P.A. Colegrove, H.E. Coules, J. Fairman, F. Martina, T. Kashoob, H. Mamash, L.D. Cozzolino, Microstructure and residual stress improvement in wire and arc additively manufactured parts through high-pressure rolling, Journal of Materials Processing Technology 213 (2013) 1782–1791. https://doi.org/10.1016/j.jmatprotec.2013.04.012

    work page doi:10.1016/j.jmatprotec.2013.04.012 2013

  17. [17]

    J. Ding, P. Colegrove, J. Mehnen, S. Ganguly, P.M. Sequeira Almeida, F. Wang, S. Williams, Thermo-mechanical analysis of Wire and Arc Additive Layer Manufacturing process on large multi-layer parts, Computational Materials Science 50 (2011) 3315–3322. https://doi.org/10.1016/j.commatsci.2011.06.023

    work page doi:10.1016/j.commatsci.2011.06.023 2011

  18. [18]

    Z. Pan, D. Ding, B. Wu, D. Cuiuri, H. Li, J. Norrish, Arc Welding Processes for Additive Manufacturing: A Review, in: S. Chen, Y. Zhang, Z. Feng (Eds.), Transactions on Intelligent Welding Manufacturing, Springer Singapore, Singapore, 2018: pp. 3–24. https://doi.org/10.1007/978-981-10-5355-9_1. 37

    work page doi:10.1007/978-981-10-5355-9_1 2018

  19. [19]

    B. Cong, J. Ding, S. Williams, Effect of arc mode in cold metal transfer process on porosity of additively manufactured Al-6.3%Cu alloy, Int J Adv Manuf Technol 76 (2015) 1593–1606. https://doi.org/10.1007/s00170-014-6346-x

    work page doi:10.1007/s00170-014-6346-x 2015

  20. [20]

    Everton, M

    S.K. Everton, M. Hirsch, P. Stravroulakis, R.K. Leach, A.T. Clare, Review of in-situ process monitoring and in-situ metrology for metal additive manufacturing, Materials & Design 95 (2016) 431–445. https://doi.org/10.1016/j.matdes.2016.01.099

    work page doi:10.1016/j.matdes.2016.01.099 2016

  21. [21]

    The strong coupling constant: state of the art and the decade ahead,

    M. Grasso, B.M. Colosimo, Process defects and in situ monitoring methods in metal powder bed fusion: a review, Meas. Sci. Technol. 28 (2017) 044005. https://doi.org/10.1088/1361- 6501/aa5c4f

    work page doi:10.1088/1361- 2017

  22. [22]

    Tapia, A

    G. Tapia, A. Elwany, A Review on Process Monitoring and Control in Metal-Based Additive Manufacturing, Journal of Manufacturing Science and Engineering 136 (2014) 060801. https://doi.org/10.1115/1.4028540

    work page doi:10.1115/1.4028540 2014

  23. [23]

    Mukherjee, T

    T. Mukherjee, T. DebRoy, Mitigation of lack of fusion defects in powder bed fusion additive manufacturing, Journal of Manufacturing Processes 36 (2018) 442–449. https://doi.org/10.1016/j.jmapro.2018.10.028

    work page doi:10.1016/j.jmapro.2018.10.028 2018

  24. [24]

    Zamiela, R

    C. Zamiela, R. Stokes, W. Tian, M.W. Priddy, L. Bian, Advancing Thermal Physics-Informed PointNet Distortion Prediction Capabilities in Wire Arc-Directed Energy Deposition, J. Comput. Inf. Sci. Eng 25 (2025). https://doi.org/10.1115/1.4069381

    work page doi:10.1115/1.4069381 2025

  25. [25]

    Zhang, Y

    Y. Zhang, Y. Chen, P. Li, A.T. Male, Weld deposition-based rapid prototyping: a preliminary study, Journal of Materials Processing Technology 135 (2003) 347–357. https://doi.org/10.1016/S0924-0136(02)00867-1

    work page doi:10.1016/s0924-0136(02)00867-1 2003

  26. [26]

    Xiong, G

    J. Xiong, G. Zhang, J. Hu, L. Wu, Bead geometry prediction for robotic GMAW-based rapid manufacturing through a neural network and a second-order regression analysis, J Intell Manuf 25 (2014) 157–163. https://doi.org/10.1007/s10845-012-0682-1

    work page doi:10.1007/s10845-012-0682-1 2014

  27. [27]

    Mattera, S.P

    G. Mattera, S.P. Chozaki, J. Norrish, Advances in machine learning for parameters optimisation and in-situ monitoring of wire arc additive manufacturing, Weld World 70 (2026) 1173–1202. https://doi.org/10.1007/s40194-025-02200-5

    work page doi:10.1007/s40194-025-02200-5 2026

  28. [28]

    Reisch, T

    R.T. Reisch, T. Hauser, B. Lutz, A. Tsakpinis, D. Winter, T. Kamps, A. Knoll, Context awareness in process monitoring of additive manufacturing using a digital twin, Int J Adv Manuf Technol 119 (2022) 3483–3500. https://doi.org/10.1007/s00170-021-08636-5

    work page doi:10.1007/s00170-021-08636-5 2022

  29. [29]

    Xiong, Y

    J. Xiong, Y. Zhang, Y. Pi, Control of deposition height in WAAM using visual inspection of previous and current layers, J Intell Manuf 32 (2021) 2209–2217. https://doi.org/10.1007/s10845- 020-01634-6

    work page doi:10.1007/s10845- 2021

  30. [30]

    B. Wu, D. Ding, Z. Pan, D. Cuiuri, H. Li, J. Han, Z. Fei, Effects of heat accumulation on the arc characteristics and metal transfer behavior in Wire Arc Additive Manufacturing of Ti6Al4V, Journal of Materials Processing Technology 250 (2017) 304–312. https://doi.org/10.1016/j.jmatprotec.2017.07.037

    work page doi:10.1016/j.jmatprotec.2017.07.037 2017

  31. [31]

    Mattera, J

    G. Mattera, J. Polden, L. Nele, Monitoring Wire Arc Additive Manufacturing process of Inconel 718 thin-walled structure using wavelet decomposition and clustering analysis of welding signal, Journal of Advanced Manufacturing Science and Technology 0 (2025) 2025006–0. https://doi.org/10.51393/j.jamst.2025006

    work page doi:10.51393/j.jamst.2025006 2025

  32. [32]

    Pringle, S

    A.M. Pringle, S. Oberloier, A.L. Petsiuk, P.G. Sanders, J.M. Pearce, Open source arc analyzer: Multi-sensor monitoring of wire arc additive manufacturing, HardwareX 8 (2020) e00137. https://doi.org/10.1016/j.ohx.2020.e00137

    work page doi:10.1016/j.ohx.2020.e00137 2020

  33. [33]

    Q. Gao, F. Lyu, L. Wang, X. Zhan, Investigation of Arc Stability in Wire Arc Additive Manufacturing of 2319 Aluminum Alloy, Materials 17 (2024) 5160. https://doi.org/10.3390/ma17215160

    work page doi:10.3390/ma17215160 2024

  34. [34]

    B. Shen, J. Lu, Y. Wang, D. Chen, J. Han, Y. Zhang, Z. Zhao, Multimodal-based weld reinforcement monitoring system for wire arc additive manufacturing, Journal of Materials Research and Technology 20 (2022) 561–571. https://doi.org/10.1016/j.jmrt.2022.07.086. 38

    work page doi:10.1016/j.jmrt.2022.07.086 2022

  35. [35]

    Huang, C

    Y. Huang, C. Yue, X. Tan, Z. Zhou, X. Li, X. Zhang, C. Zhou, Y. Peng, K. Wang, Quality Prediction for Wire Arc Additive Manufacturing Based on Multi-source Signals, Whale Optimization Algorithm–Variational Modal Decomposition, and One-Dimensional Convolutional Neural Network, J. of Materi Eng and Perform 33 (2024) 11351–11364. https://doi.org/10.1007/s116...

    work page doi:10.1007/s11665-023-08768-7 2024

  36. [36]

    A. Love, S. Behseresht, Y.H. Park, Reinforcement Learning Enabled Intelligent Process Monitoring and Control of Wire Arc Additive Manufacturing, JMMP 9 (2025) 340. https://doi.org/10.3390/jmmp9100340

    work page doi:10.3390/jmmp9100340 2025

  37. [37]

    Afrasiabi, M

    N.A. Surovi, G.S. Soh, Acoustic feature based geometric defect identification in wire arc additive manufacturing, Virtual and Physical Prototyping 18 (2023) e2210553. https://doi.org/10.1080/17452759.2023.2210553

    work page doi:10.1080/17452759.2023.2210553 2023

  38. [38]

    Q. Zhu, Z. Huang, H. Li, Research Progress of Acoustic Monitoring Technology in Welding and Additive Manufacturing Processes, Micromachines 17 (2026) 246. https://doi.org/10.3390/mi17020246

    work page doi:10.3390/mi17020246 2026

  39. [39]

    Pal, S.K

    K. Pal, S.K. Pal, Effect of Pulse Parameters on Weld Quality in Pulsed Gas Metal Arc Welding: A Review, J. of Materi Eng and Perform 20 (2011) 918–931. https://doi.org/10.1007/s11665-010- 9717-y

    work page doi:10.1007/s11665-010- 2011

  40. [40]

    Thomas, Analyzing the Failure of Welded Steel Components in Construction Systems, J Fail

    D.J. Thomas, Analyzing the Failure of Welded Steel Components in Construction Systems, J Fail. Anal. and Preven. 18 (2018) 304–314. https://doi.org/10.1007/s11668-018-0392-x

    work page doi:10.1007/s11668-018-0392-x 2018

  41. [41]

    Rahman, S

    M.A. Rahman, S. Jamal, M.V. Cruz, B. Silwal, H. Taheri, In situ process monitoring of multi-layer deposition in wire arc additive manufacturing (WAAM) process with acoustic data analysis and machine learning, Int J Adv Manuf Technol 132 (2024) 5087–5101. https://doi.org/10.1007/s00170-024-13641-5

    work page doi:10.1007/s00170-024-13641-5 2024

  42. [42]

    https://insidemetaladditivemanufacturing.com/2024/02/21/acoustic-emission-monitoring-in- selective-laser-melting-technique-definition-and-accuracy/ (accessed April 9, 2026)

    Acoustic Emission Monitoring in Selective Laser Melting: Technique Definition and Accuracy – Inside Metal Additive Manufacturing, (n.d.). https://insidemetaladditivemanufacturing.com/2024/02/21/acoustic-emission-monitoring-in- selective-laser-melting-technique-definition-and-accuracy/ (accessed April 9, 2026)

    2024

  43. [43]

    Stemmer, J.A

    G. Stemmer, J.A. Lopez, J.A.D.H. Ontiveros, A. Raju, T. Thimmanaik, S. Biswas, Unsupervised Welding Defect Detection Using Audio And Video, (2024). https://doi.org/10.48550/ARXIV.2409.02290

    work page doi:10.48550/arxiv.2409.02290 2024

  44. [44]

    Sumesh, K

    A. Sumesh, K. Rameshkumar, K. Mohandas, R.S. Babu, Use of Machine Learning Algorithms for Weld Quality Monitoring using Acoustic Signature, Procedia Computer Science 50 (2015) 316–

    2015

  45. [45]

    https://doi.org/10.1016/j.procs.2015.04.042

    work page doi:10.1016/j.procs.2015.04.042 2015

  46. [46]

    Zhang, Q

    H. Zhang, Q. Wu, W. Tang, J. Yang, Acoustic Signal-Based Defect Identification for Directed Energy Deposition-Arc Using Wavelet Time–Frequency Diagrams, Sensors 24 (2024) 4397. https://doi.org/10.3390/s24134397

    work page doi:10.3390/s24134397 2024

  47. [47]

    W. Li, H. Zhang, G. Wang, G. Xiong, M. Zhao, G. Li, R. Li, Deep learning based online metallic surface defect detection method for wire and arc additive manufacturing, Robotics and Computer- Integrated Manufacturing 80 (2023) 102470. https://doi.org/10.1016/j.rcim.2022.102470

    work page doi:10.1016/j.rcim.2022.102470 2023

  48. [48]

    Mattera, L

    G. Mattera, L. Nele, Machine learning approaches for real-time process anomaly detection in wire arc additive manufacturing, Int J Adv Manuf Technol 137 (2025) 2863–2888. https://doi.org/10.1007/s00170-025-15327-y

    work page doi:10.1007/s00170-025-15327-y 2025

  49. [49]

    Thompson, I

    A. Thompson, I. Maskery, R.K. Leach, X-ray computed tomography for additive manufacturing: a review, Meas. Sci. Technol. 27 (2016) 072001. https://doi.org/10.1088/0957-0233/27/7/072001

    work page doi:10.1088/0957-0233/27/7/072001 2016

  50. [50]

    Du Plessis, I

    A. Du Plessis, I. Yadroitsev, I. Yadroitsava, S.G. Le Roux, X-Ray Microcomputed Tomography in Additive Manufacturing: A Review of the Current Technology and Applications, 3D Printing and Additive Manufacturing 5 (2018) 227–247. https://doi.org/10.1089/3dp.2018.0060

    work page doi:10.1089/3dp.2018.0060 2018

  51. [51]

    Bellens, P

    S. Bellens, P. Vandewalle, W. Dewulf, Deep learning based porosity segmentation in X-ray CT measurements of polymer additive manufacturing parts, Procedia CIRP 96 (2021) 336–341. https://doi.org/10.1016/j.procir.2021.01.157. 39

    work page doi:10.1016/j.procir.2021.01.157 2021

  52. [52]

    Gobert, A

    C. Gobert, A. Kudzal, J. Sietins, C. Mock, J. Sun, B. McWilliams, Porosity segmentation in X-ray computed tomography scans of metal additively manufactured specimens with machine learning, Additive Manufacturing 36 (2020) 101460. https://doi.org/10.1016/j.addma.2020.101460

    work page doi:10.1016/j.addma.2020.101460 2020

  53. [53]

    Kantzos, J

    C. Kantzos, J. Pauza, R. Cunningham, S.P. Narra, J. Beuth, A. Rollett, An Investigation of Process Parameter Modifications on Additively Manufactured Inconel 718 Parts, J. of Materi Eng and Perform 28 (2019) 620–626. https://doi.org/10.1007/s11665-018-3612-3

    work page doi:10.1007/s11665-018-3612-3 2019

  54. [54]

    Seifi, A

    M. Seifi, A. Salem, J. Beuth, O. Harrysson, J.J. Lewandowski, Overview of Materials Qualification Needs for Metal Additive Manufacturing, JOM 68 (2016) 747–764. https://doi.org/10.1007/s11837-015-1810-0

    work page doi:10.1007/s11837-015-1810-0 2016

  55. [55]

    Slotwinski, E.J

    J.A. Slotwinski, E.J. Garboczi, P.E. Stutzman, C.F. Ferraris, S.S. Watson, M.A. Peltz, Characterization of Metal Powders Used for Additive Manufacturing, J. RES. NATL. INST. STAN. 119 (2014) 460. https://doi.org/10.6028/jres.119.018

    work page doi:10.6028/jres.119.018 2014

  56. [56]

    Scime, J

    L. Scime, J. Beuth, Anomaly detection and classification in a laser powder bed additive manufacturing process using a trained computer vision algorithm, Additive Manufacturing 19 (2018) 114–126. https://doi.org/10.1016/j.addma.2017.11.009

    work page doi:10.1016/j.addma.2017.11.009 2018

  57. [57]

    Wilbig, A.E

    J. Wilbig, A.E. Wilson-Heid, L. Bernard, J. Baptista, A.-F. Obaton, Comparison of Porosity Analysis Based on X-Ray Computed Tomography on Metal Parts Produced by Additive Manufacturing, Applied Sciences 15 (2025) 9876. https://doi.org/10.3390/app15189876

    work page doi:10.3390/app15189876 2025

  58. [58]

    Grasso, A.G

    M. Grasso, A.G. Demir, B. Previtali, B.M. Colosimo, In situ monitoring of selective laser melting of zinc powder via infrared imaging of the process plume, Robotics and Computer-Integrated Manufacturing 49 (2018) 229–239. https://doi.org/10.1016/j.rcim.2017.07.001

    work page doi:10.1016/j.rcim.2017.07.001 2018

  59. [59]

    W.L. Ng, G.L. Goh, G.D. Goh, J.S.J. Ten, W.Y. Yeong, Progress and Opportunities for Machine Learning in Materials and Processes of Additive Manufacturing, Advanced Materials 36 (2024) 2310006. https://doi.org/10.1002/adma.202310006

    work page doi:10.1002/adma.202310006 2024

  60. [60]

    Deshpande, V

    S. Deshpande, V. Venugopal, M. Kumar, S. Anand, Deep learning-based image segmentation for defect detection in additive manufacturing: an overview, Int J Adv Manuf Technol 134 (2024) 2081–2105. https://doi.org/10.1007/s00170-024-14191-6

    work page doi:10.1007/s00170-024-14191-6 2024

  61. [61]

    Kunkel, D

    N. Kunkel, D. Thölken, K. Behler, Deep learning-based automated defect classification for powder bed fusion – Laser beam, European Journal of Materials 4 (2024) 2427401. https://doi.org/10.1080/26889277.2024.2427401

    work page doi:10.1080/26889277.2024.2427401 2024

  62. [62]

    Saimon, E

    A.I. Saimon, E. Yangue, X. Yue, Z.J. Kong, C. Liu, Advancing Additive Manufacturing through Deep Learning: A Comprehensive Review of Current Progress and Future Challenges, (2024). https://doi.org/10.48550/ARXIV.2403.00669

    work page doi:10.48550/arxiv.2403.00669 2024

  63. [63]

    Y. Mao, H. Lin, C.X. Yu, R. Frye, D. Beckett, K. Anderson, L. Jacquemetton, F. Carter, Z. Gao, W. Liao, A.N. Choudhary, K. Ehmann, A. Agrawal, A deep learning framework for layer-wise porosity prediction in metal powder bed fusion using thermal signatures, J Intell Manuf 34 (2023) 315–329. https://doi.org/10.1007/s10845-022-02039-3

    work page doi:10.1007/s10845-022-02039-3 2023

  64. [64]

    Caggiano, J

    A. Caggiano, J. Zhang, V. Alfieri, F. Caiazzo, R. Gao, R. Teti, Machine learning-based image processing for on-line defect recognition in additive manufacturing, CIRP Annals 68 (2019) 451–

    2019

  65. [65]

    https://doi.org/10.1016/j.cirp.2019.03.021

    work page doi:10.1016/j.cirp.2019.03.021 2019

  66. [66]

    K.W. Kim, A. Kamerkar, T.-E. Chiu, I. Abdi, J. Qin, W. Suder, S. Asif, WAAM-ViD: towards universal vision-based monitoring for wire arc additive manufacturing, Front. Manuf. Technol. 5 (2025) 1676365. https://doi.org/10.3389/fmtec.2025.1676365

    work page doi:10.3389/fmtec.2025.1676365 2025

  67. [67]

    Alcaraz, A

    J.Y. Alcaraz, A. Sharma, T. Tjahjowidodo, Predicting porosity in wire arc additive manufacturing (WAAM) using wavelet scattering networks and sparse principal component analysis, Weld World 68 (2024) 843–853. https://doi.org/10.1007/s40194-024-01709-5

    work page doi:10.1007/s40194-024-01709-5 2024

  68. [68]

    Nadiu, V

    J.B. Nadiu, V. Sushma, J.P. C.T, M. Dhanush, V. Senthilkumar, A deep learning-based model for defect reorganisation in welding/wire arc additive manufacturing, Welding International 39 (2025) 76–85. https://doi.org/10.1080/09507116.2024.2439889. 40

    work page doi:10.1080/09507116.2024.2439889 2025

  69. [69]

    Y. Li, J. Polden, Z. Pan, J. Cui, C. Xia, F. He, H. Mu, H. Li, L. Wang, A defect detection system for wire arc additive manufacturing using incremental learning, Journal of Industrial Information Integration 27 (2022) 100291. https://doi.org/10.1016/j.jii.2021.100291

    work page doi:10.1016/j.jii.2021.100291 2022

  70. [70]

    Language Models are Few-Shot Learners

    T.B. Brown, B. Mann, N. Ryder, M. Subbiah, I. Sutskever, D. Amodei, Language Models are Few-Shot Learners, (2020). https://doi.org/10.48550/ARXIV.2005.14165

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2005.14165 2020

  71. [71]

    Attention Is All You Need

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A.N. Gomez, L. Kaiser, I. Polosukhin, Attention Is All You Need, (2023). https://doi.org/10.48550/arXiv.1706.03762

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1706.03762 2023

  72. [72]

    LLaMA: Open and Efficient Foundation Language Models

    H. Touvron, T. Lavril, G. Izacard, X. Martinet, M.-A. Lachaux, T. Lacroix, B. Rozière, N. Goyal, E. Hambro, F. Azhar, A. Rodriguez, A. Joulin, E. Grave, G. Lample, LLaMA: Open and Efficient Foundation Language Models, (2023). https://doi.org/10.48550/ARXIV.2302.13971

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2302.13971 2023

  73. [73]

    GPT-4 Technical Report

    OpenAI, J. Achiam, S. Adler, S. Agarwal, W. Zhuk, B. Zoph, GPT-4 Technical Report, (2023). https://doi.org/10.48550/ARXIV.2303.08774

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2303.08774 2023

  74. [74]

    J. Wei, X. Wang, D. Schuurmans, M. Bosma, B. Ichter, F. Xia, E. Chi, Q. Le, D. Zhou, Chain-of- Thought Prompting Elicits Reasoning in Large Language Models, (2022). https://doi.org/10.48550/ARXIV.2201.11903

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2201.11903 2022

  75. [75]

    Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks

    P. Lewis, E. Perez, A. Piktus, F. Petroni, V. Karpukhin, N. Goyal, H. Küttler, M. Lewis, W. Yih, T. Rocktäschel, S. Riedel, D. Kiela, Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks, (2020). https://doi.org/10.48550/ARXIV.2005.11401

    work page internal anchor Pith review doi:10.48550/arxiv.2005.11401 2020

  76. [76]

    A. M. Bran, S. Cox, O. Schilter, C. Baldassari, A.D. White, P. Schwaller, Augmenting large language models with chemistry tools, Nat Mach Intell 6 (2024) 525–535. https://doi.org/10.1038/s42256-024-00832-8

    work page doi:10.1038/s42256-024-00832-8 2024

  77. [77]

    Chandrasekhar, J

    A. Chandrasekhar, J. Chan, F. Ogoke, O. Ajenifujah, A. Barati Farimani, AMGPT: A large language model for contextual querying in additive manufacturing, Additive Manufacturing Letters 11 (2024) 100232. https://doi.org/10.1016/j.addlet.2024.100232

    work page doi:10.1016/j.addlet.2024.100232 2024

  78. [78]

    Jignasu, K

    A. Jignasu, K. Marshall, B. Ganapathysubramanian, A. Balu, C. Hegde, A. Krishnamurthy, Towards Foundational AI Models for Additive Manufacturing: Language Models for G-Code Debugging, Manipulation, and Comprehension, (2023). https://doi.org/10.48550/ARXIV.2309.02465

    work page doi:10.48550/arxiv.2309.02465 2023

  79. [79]

    Galdino, T

    M. Galdino, T. Hamann, A. Abdelrazeq, I. Isenhardt, Large Language Model‐Based Cognitive Assistants for Quality Management Systems in Manufacturing: A Requirement Analysis, Engineering Reports 7 (2025) e70437. https://doi.org/10.1002/eng2.70437

    work page doi:10.1002/eng2.70437 2025

  80. [80]

    Huang, Y

    J. Huang, Y. Teng, Q. Liu, L. Gao, X. Li, C. Zhang, G. Xu, Leveraging large language models for efficient scheduling in Human–Robot collaborative flexible manufacturing systems, Npj Adv. Manuf. 2 (2025) 47. https://doi.org/10.1038/s44334-025-00061-w

    work page doi:10.1038/s44334-025-00061-w 2025

Showing first 80 references.