62470 2025 eng 1 13 286 article Elsevier B.V. 1 -- -- -- The laser welding process is an important manufacturing technology for metallic materials. However, its application is often hindered by the occurrence of porosity defects. By far, an accurate prediction of the porosity defects and an insight into its formation mechanism are still challenging due to the highly nonlinear physics involved. In this paper, we propose a physics-informed deep learning (PIDL) framework by utilizing mechanistic modeling and experimental data to predict the porosity level during laser beam welding of aluminum alloys. With a proper selection of the physical variables (features) concerning the solidification, liquid metal flow, keyhole stability, and weld pool geometry, the PIDL model shows great superiority in predicting the porosity ratio, with a reduction of mean square error by 41 %, in comparison with the conventional DL model trained with welding parameters. Furthermore, the selected variables are fused into dimensionless features with explicit physical meanings to improve the interpretability and extendibility of the PIDL model. Based on a well-trained PIDL model, the hierarchical importance of the physical variables/procedures on the porosity formation is for the first time revealed with the help of the Shapley Additive Explanations analysis. The keyhole ratio is identified as the most influential factor in the porosity formation, followed by the downward flow-driven drag force, which offers a valuable guideline for process optimization and porosity minimization. Acta Materialia 10.1016/j.actamat.2025.120740 urn:nbn:de:kobv:b43-624708 true {"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,1,28]],"date-time":"2025-01-28T05:22:49Z","timestamp":1738041769095,"version":"3.33.0"},"reference-count":51,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2025,3,1]],"date-time":"2025-03-01T00:00:00Z","timestamp":1740787200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2025,3,1]],"date-time":"2025-03-01T00:00:00Z","timestamp":1740787200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/legal\/tdmrep-license"},{"start":{"date-parts":[[2025,1,11]],"date-time":"2025-01-11T00:00:00Z","timestamp":1736553600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001659","name":"German Research Foundation","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Acta Materialia"],"published-print":{"date-parts":[[2025,3]]},"DOI":"10.1016\/j.actamat.2025.120740","type":"journal-article","created":{"date-parts":[[2025,1,11]],"date-time":"2025-01-11T23:23:24Z","timestamp":1736637804000},"page":"120740","update-policy":"https:\/\/doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":0,"special_numbering":"C","title":["Toward prediction and insight of porosity formation in laser welding: A physics-informed deep learning framework"],"prefix":"10.1016","volume":"286","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8996-759X","authenticated-orcid":false,"given":"Xiangmeng","family":"Meng","sequence":"first","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2395-6462","authenticated-orcid":false,"given":"Marcel","family":"Bachmann","sequence":"additional","affiliation":[]},{"given":"Fan","family":"Yang","sequence":"additional","affiliation":[]},{"given":"Michael","family":"Rethmeier","sequence":"additional","affiliation":[]}],"member":"78","reference":[{"key":"10.1016\/j.actamat.2025.120740_bib0001","doi-asserted-by":"crossref","first-page":"43","DOI":"10.1080\/13621718.2021.1999763","article-title":"High-power laser beam welding for thick section steels \u2013 new perspectives using electromagnetic systems","volume":"27","author":"Rethmeier","year":"2022","journal-title":"Sci. Technol. Weld. Join."},{"key":"10.1016\/j.actamat.2025.120740_bib0002","first-page":"419","article-title":"The effect of reduced pressure on laser keyhole weld porosity and weld geometry in commercially pure titanium and nickel","volume":"95","author":"Elmer","year":"2016","journal-title":"Weld. J."},{"key":"10.1016\/j.actamat.2025.120740_bib0003","doi-asserted-by":"crossref","first-page":"251","DOI":"10.1016\/j.actamat.2016.12.062","article-title":"Keyhole formation and thermal fluid flow-induced porosity during laser fusion welding in titanium alloys: experimental and modelling","volume":"126","author":"Panwisawas","year":"2017","journal-title":"Acta Mater"},{"key":"10.1016\/j.actamat.2025.120740_bib0004","doi-asserted-by":"crossref","first-page":"206","DOI":"10.1080\/09507116.2016.1223204","article-title":"Elucidation of the effect of welding speed on melt flows in high-brightness and high-power laser welding of stainless steel on basis of three-dimensional X-ray transmission in situ observation","volume":"31","author":"Kawahito","year":"2017","journal-title":"Weld. Int."},{"key":"10.1016\/j.actamat.2025.120740_bib0005","doi-asserted-by":"crossref","first-page":"2808","DOI":"10.1007\/s11661-014-2231-3","article-title":"A quantitative model of keyhole instability induced porosity in laser welding of titanium alloy","volume":"45","author":"Pang","year":"2014","journal-title":"Metall. Mater. Trans. A"},{"key":"10.1016\/j.actamat.2025.120740_bib0006","doi-asserted-by":"crossref","first-page":"1143","DOI":"10.1016\/j.ijheatmasstransfer.2015.07.041","article-title":"Formation and influence mechanism of keyhole-induced porosity in deep-penetration laser welding based on 3D transient modeling","volume":"90","author":"Lu","year":"2015","journal-title":"Int. J. Heat. Mass. Transf."},{"key":"10.1016\/j.actamat.2025.120740_bib0007","doi-asserted-by":"crossref","DOI":"10.1016\/j.optlastec.2020.106562","article-title":"Geometry and stability of the capillary during deep-penetration laser welding of AlMgSi at high feed rates","volume":"133","author":"Fetzer","year":"2021","journal-title":"Opt. Laser. Technol."},{"key":"10.1016\/j.actamat.2025.120740_bib0008","doi-asserted-by":"crossref","first-page":"9","DOI":"10.1016\/j.jmatprotec.2017.06.033","article-title":"Dynamics of solid-liquid interface and porosity formation determined through x-ray phase-contrast in laser welding of pure Al","volume":"250","author":"Miyagi","year":"2017","journal-title":"J. Mater. Process. Technol."},{"key":"10.1016\/j.actamat.2025.120740_bib0009","doi-asserted-by":"crossref","first-page":"1409","DOI":"10.1007\/s11837-012-0481-3","article-title":"Effect of travel speed and beam focus on porosity in alloy 690 laser welds","volume":"64","author":"Tucker","year":"2012","journal-title":"JOM"},{"key":"10.1016\/j.actamat.2025.120740_bib0010","doi-asserted-by":"crossref","first-page":"452","DOI":"10.3390\/met7110452","article-title":"Control of porosity and spatter in laser welding of thick AlMg5 Parts using high-speed imaging and optical microscopy","volume":"7","author":"Popescu","year":"2017","journal-title":"Metals (Basel)"},{"key":"10.1016\/j.actamat.2025.120740_bib0011","first-page":"165","article-title":"Thermographic detectability of porosity in multi-layered bare and anodized aluminum laser welds","volume":"48","author":"Rinker","year":"2020","journal-title":"Proced. Manuf."},{"key":"10.1016\/j.actamat.2025.120740_bib0012","article-title":"Monitoring of YAG laser spot welding-detection of porosity defect by acoustic signal","volume":"1","author":"Sasaki","year":"2002","journal-title":"Int. Congr. Appl. Lasers Electro-Opt."},{"key":"10.1016\/j.actamat.2025.120740_bib0013","first-page":"62","article-title":"Deep-learning-based porosity monitoring of laser welding process","volume":"23","author":"Zhang","year":"2020","journal-title":"Manuf. Lett."},{"key":"10.1016\/j.actamat.2025.120740_bib0014","article-title":"Online porosity defect detection based on convolutional neural network for Al alloy laser welding","volume":"1884","author":"Ma","year":"2021","journal-title":"J.Phys.: Conf. Ser."},{"key":"10.1016\/j.actamat.2025.120740_bib0015","doi-asserted-by":"crossref","first-page":"1687","DOI":"10.1007\/s10845-023-02117-0","article-title":"A deep learning framework for defect prediction based on thermographic in-situ monitoring in laser powder bed fusion","volume":"35","author":"Oster","year":"2024","journal-title":"J. Intell. Manuf."},{"key":"10.1016\/j.actamat.2025.120740_bib0016","doi-asserted-by":"crossref","first-page":"19493","DOI":"10.1038\/s41598-020-75131-4","article-title":"In-situ porosity recognition for laser additive manufacturing of 7075-Al alloy using plasma emission spectroscopy","volume":"10","author":"Ren","year":"2020","journal-title":"Sci. Rep."},{"key":"10.1016\/j.actamat.2025.120740_bib0017","doi-asserted-by":"crossref","first-page":"2719","DOI":"10.1007\/s10845-023-02170-9","article-title":"In-situ porosity prediction in metal powder bed fusion additive manufacturing using spectral emissions: a prior-guided machine learning approach","volume":"35","author":"Atwya","year":"2023","journal-title":"J. Intell. Manuf."},{"key":"10.1016\/j.actamat.2025.120740_bib0018","doi-asserted-by":"crossref","first-page":"561","DOI":"10.1007\/s00170-016-9366-x","article-title":"Defects monitoring of laser metal deposition using acoustic emission sensor","volume":"90","author":"Gaja","year":"2017","journal-title":"Int. J. Adv. Manuf. Technol."},{"key":"10.1016\/j.actamat.2025.120740_bib0019","article-title":"Correlations between thermal history and keyhole porosity in laser powder bed fusion","volume":"34","author":"Paulson","year":"2020","journal-title":"Addit. Manuf."},{"key":"10.1016\/j.actamat.2025.120740_bib0020","doi-asserted-by":"crossref","first-page":"334","DOI":"10.1016\/j.jmsy.2021.12.004","article-title":"Multi-sensing signals diagnosis and CNN-based detection of porosity defect during Al alloys laser welding","volume":"62","author":"Ma","year":"2022","journal-title":"J. Manuf. Sys."},{"key":"10.1016\/j.actamat.2025.120740_bib0021","doi-asserted-by":"crossref","first-page":"70","DOI":"10.1016\/j.jmsy.2022.08.011","article-title":"Real-time porosity monitoring during laser welding of aluminum alloys based on keyhole 3D morphology characteristics","volume":"65","author":"Ma","year":"2022","journal-title":"J. Manuf. Sys."},{"key":"10.1016\/j.actamat.2025.120740_bib0022","doi-asserted-by":"crossref","first-page":"499","DOI":"10.1007\/s00466-022-02257-9","article-title":"Hybrid thermal modeling of additive manufacturing processes using physics-informed neural networks for temperature prediction and parameter identification","volume":"72","author":"Liao","year":"2023","journal-title":"Comput. Mech."},{"issue":"2","key":"10.1016\/j.actamat.2025.120740_bib0023","doi-asserted-by":"crossref","first-page":"619","DOI":"10.1007\/s00466-020-01952-9","article-title":"Machine learning for metal additive manufacturing: predicting temperature and melt pool fluid dynamics using physics-informed neural networks","volume":"67","author":"Zhu","year":"2021","journal-title":"Comput. Mech."},{"key":"10.1016\/j.actamat.2025.120740_bib0024","doi-asserted-by":"crossref","first-page":"67","DOI":"10.1016\/j.actamat.2020.03.047","article-title":"Machine learning based hierarchy of causative variables for tool failure in friction stir welding","volume":"192","author":"Du","year":"2020","journal-title":"Acta Mater"},{"key":"10.1016\/j.actamat.2025.120740_bib0025","doi-asserted-by":"crossref","DOI":"10.1016\/j.actamat.2021.117612","article-title":"Crack free metal printing using physics informed machine learning","volume":"226","author":"Mondal","year":"2022","journal-title":"Acta Mater"},{"key":"10.1016\/j.actamat.2025.120740_bib0026","doi-asserted-by":"crossref","first-page":"875","DOI":"10.1016\/j.jmsy.2021.07.013","article-title":"Leveraging simulated and empirical data-driven insight to supervised-learning for porosity prediction in laser metal deposition","volume":"62","author":"Gawade","year":"2022","journal-title":"J. Manuf. Sys."},{"key":"10.1016\/j.actamat.2025.120740_bib0027","doi-asserted-by":"crossref","first-page":"55","DOI":"10.1007\/s10845-022-02033-9","article-title":"Online porosity prediction in laser welding of aluminum alloys based on a multi-fidelity deep learning framework","volume":"35","author":"Ma","year":"2024","journal-title":"J. Intell. Manuf."},{"key":"10.1016\/j.actamat.2025.120740_bib0028","doi-asserted-by":"crossref","first-page":"59","DOI":"10.1016\/j.apmt.2018.11.003","article-title":"A digital twin for rapid qualification of 3D printed metallic components","volume":"14","author":"Mukherjee","year":"2019","journal-title":"Appl. Mater. Today"},{"key":"10.1016\/j.actamat.2025.120740_bib0029","doi-asserted-by":"crossref","DOI":"10.1016\/j.jmatprotec.2023.118265","article-title":"A fundamental study of physical mechanisms of wineglass-shaped fusion zone profile in laser melting","volume":"324","author":"Meng","year":"2024","journal-title":"J. Mater. Process. Technol."},{"key":"10.1016\/j.actamat.2025.120740_bib0030","doi-asserted-by":"crossref","DOI":"10.2351\/7.0001078","article-title":"Influence of the spatial laser energy absorption on the molten pool dynamics in high-power laser beam welding","volume":"35","author":"Meng","year":"2023","journal-title":"J. Laser. Appl."},{"key":"10.1016\/j.actamat.2025.120740_bib0031","doi-asserted-by":"crossref","first-page":"2541","DOI":"10.1088\/0022-3727\/30\/18\/008","article-title":"The role of recoil pressure in energy balance during laser materials processing","volume":"30","author":"Vladimir","year":"1997","journal-title":"J. Phys. D: Appl. Phys."},{"key":"10.1016\/j.actamat.2025.120740_bib0032","doi-asserted-by":"crossref","first-page":"645","DOI":"10.1016\/j.ijheatmasstransfer.2009.10.030","article-title":"A coupled volume-of-fluid and level set (VOSET) method for computing incompressible two-phase flows","volume":"53","author":"Sun","year":"2010","journal-title":"Int. J.Heat. Mass. Transf."},{"key":"10.1016\/j.actamat.2025.120740_bib0033","doi-asserted-by":"crossref","first-page":"5372","DOI":"10.1088\/0022-3727\/39\/24\/039","article-title":"Implementation of real-time multiple reflection and Fresnel absorption of laser beam in keyhole","volume":"39","author":"Jung-Ho","year":"2006","journal-title":"J. Phys. D: Appl. Phys."},{"key":"10.1016\/j.actamat.2025.120740_bib0034","doi-asserted-by":"crossref","first-page":"262","DOI":"10.1016\/j.jmatprotec.2011.09.011","article-title":"Numerical simulation of molten pool dynamics in high power disk laser welding","volume":"212","author":"Cho","year":"2012","journal-title":"J. Mater. Process. Technol."},{"key":"10.1016\/j.actamat.2025.120740_bib0035","doi-asserted-by":"crossref","DOI":"10.1088\/0022-3727\/44\/2\/025301","article-title":"A three-dimensional sharp interface model for self-consistent keyhole and weld pool dynamics in deep penetration laser welding","volume":"44","author":"Pang","year":"2011","journal-title":"J. Phys. D: Appl. Phys."},{"key":"10.1016\/j.actamat.2025.120740_bib0036","doi-asserted-by":"crossref","first-page":"11","DOI":"10.1016\/j.phpro.2011.03.003","article-title":"Numerical simulations - a versatile approach for better understanding dynamics in laser material processing","volume":"12","author":"Otto","year":"2011","journal-title":"Phys. Proced."},{"issue":"9","key":"10.1016\/j.actamat.2025.120740_bib0037","doi-asserted-by":"crossref","DOI":"10.1002\/adem.202100137","article-title":"Modeling laser beam absorption of metal alloys at high temperatures for selective laser melting","volume":"23","author":"Yang","year":"2021","journal-title":"Adv. Eng. Mater."},{"key":"10.1016\/j.actamat.2025.120740_bib0038","article-title":"Bulk-explosion-induced metal spattering during laser processing","volume":"9","author":"Zhao","year":"2019","journal-title":"Phys. Rev. X"},{"key":"10.1016\/j.actamat.2025.120740_bib0039","doi-asserted-by":"crossref","first-page":"849","DOI":"10.1126\/science.aav4687","article-title":"Keyhole threshold and morphology in laser melting revealed by ultrahigh-speed x-ray imaging","volume":"363","author":"Cunningham","year":"2019","journal-title":"Sci"},{"key":"10.1016\/j.actamat.2025.120740_bib0040","doi-asserted-by":"crossref","first-page":"22","DOI":"10.1038\/s41524-022-00699-6","article-title":"Mechanism of keyhole pore formation in metal additive manufacturing","volume":"8","author":"Wang","year":"2022","journal-title":"npj Comput. Mater."},{"key":"10.1016\/j.actamat.2025.120740_bib0041","doi-asserted-by":"crossref","first-page":"2379","DOI":"10.1038\/s41467-021-22704-0","article-title":"Universal scaling laws of keyhole stability and porosity in 3D printing of metals","volume":"12","author":"Gan","year":"2021","journal-title":"Nat. Commun."},{"key":"10.1016\/j.actamat.2025.120740_bib0042","doi-asserted-by":"crossref","first-page":"1563","DOI":"10.1016\/j.jmatprotec.2011.04.002","article-title":"Relationship of laser absorption to keyhole behavior in high power fiber laser welding of stainless steel and aluminum alloy","volume":"211","author":"Kawahito","year":"2011","journal-title":"J. Mater. Process. Technol."},{"key":"10.1016\/j.actamat.2025.120740_bib0043","doi-asserted-by":"crossref","DOI":"10.1016\/j.ijheatmasstransfer.2022.123084","article-title":"Numerical study of keyhole dynamics and porosity formation during high-power oscillating laser welding of medium-thick aluminum alloy plates","volume":"194","author":"Geng","year":"2022","journal-title":"Int. J. Heat. Mass. Transf."},{"key":"10.1016\/j.actamat.2025.120740_bib0044","series-title":"31st Conference on Neural Information Processing Systems","article-title":"A unified approach to interpreting model predictions","author":"Lundberg","year":"2017"},{"key":"10.1016\/j.actamat.2025.120740_bib0045","doi-asserted-by":"crossref","first-page":"184","DOI":"10.1038\/s41524-022-00866-9","article-title":"Superior printed parts using history and augmented machine learning","volume":"8","author":"Jiang","year":"2022","journal-title":"npj Comput. Mater."},{"key":"10.1016\/j.actamat.2025.120740_bib0046","article-title":"Physics-informed machine learning and mechanistic modeling of additive manufacuring to reduce defects","volume":"24","author":"Du","year":"2021","journal-title":"Appl. Mater. Today"},{"key":"10.1016\/j.actamat.2025.120740_bib0047","doi-asserted-by":"crossref","first-page":"68","DOI":"10.1038\/s41524-019-0207-y","article-title":"Conditions for void formation in friction stir welding from machine learning","volume":"5","author":"Du","year":"2019","journal-title":"npj Comput. Mater."},{"key":"10.1016\/j.actamat.2025.120740_bib0048","article-title":"Benchmark study of melt pool and keyhole dynamics, laser absorptance, and porosity in additive manufacturing. of Ti-6Al-4V","author":"Samaei","year":"2024","journal-title":"Prog. Addit. Manuf."},{"key":"10.1016\/j.actamat.2025.120740_bib0049","doi-asserted-by":"crossref","first-page":"123","DOI":"10.1007\/BF00058655","article-title":"Bagging predictors","volume":"24","author":"Breiman","year":"1996","journal-title":"Mach. Learn."},{"key":"10.1016\/j.actamat.2025.120740_bib0050","doi-asserted-by":"crossref","first-page":"481","DOI":"10.1088\/0022-3727\/26\/3\/022","article-title":"Dynamic behaviour of the keyhole in laser welding","volume":"26","author":"Kroos","year":"1993","journal-title":"J. Phys. D: Appl. Phys."},{"key":"10.1016\/j.actamat.2025.120740_bib0051","doi-asserted-by":"crossref","first-page":"44","DOI":"10.1179\/136217110X12785889549949","article-title":"High speed imaging technique Part 2 \u2013 High speed imaging of power beam welding phenomena","volume":"16","author":"Tsukamoto","year":"2011","journal-title":"Sci. Technol. Weld. Join."}],"container-title":["Acta Materialia"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S1359645425000333?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S1359645425000333?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2025,1,27]],"date-time":"2025-01-27T08:28:26Z","timestamp":1737966506000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S1359645425000333"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,3]]},"references-count":51,"alternative-id":["S1359645425000333"],"URL":"https:\/\/doi.org\/10.1016\/j.actamat.2025.120740","relation":{},"ISSN":["1359-6454"],"issn-type":[{"type":"print","value":"1359-6454"}],"subject":[],"published":{"date-parts":[[2025,3]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"Toward prediction and insight of porosity formation in laser welding: A physics-informed deep learning framework","name":"articletitle","label":"Article Title"},{"value":"Acta Materialia","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.actamat.2025.120740","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2025 The Authors. Published by Elsevier Inc. on behalf of Acta Materialia Inc.","name":"copyright","label":"Copyright"}],"article-number":"120740"}} journal-article publish 1 05.05.2025 Creative Commons - CC BY-NC-ND - Namensnennung - Nicht kommerziell - Keine Bearbeitungen 4.0 International Xiangmeng Meng Marcel Bachmann Fan Yang Michael Rethmeier eng uncontrolled Laser beam welding eng uncontrolled Physics-informed deep learning eng uncontrolled Porosity prediction eng uncontrolled Feature fusion eng uncontrolled Hierarchical importance Ingenieurwissenschaften und zugeordnete Tätigkeiten 9 Komponentensicherheit 9.3 Schweißtechnische Fertigungsverfahren Material Degradationsmechanismen Verlagsliteratur Datei für die Öffentlichkeit verfügbar ("Open Access") Wissenschaftliche Artikel der BAM Bundesanstalt für Materialforschung und -prüfung (BAM) https://opus4.kobv.de/opus4-bam/files/62470/1-s2.0-S1359645425000333-main.pdf