{"title":"基于声发射的LPBF不稳定熔池状态的不确定性驱动可信识别范式","authors":"Jiafeng Tang , Kunpeng Tan , Junlong Tang , Zhibin Zhao , Xingwu Zhang , Xuefeng Chen","doi":"10.1016/j.addma.2025.104887","DOIUrl":null,"url":null,"abstract":"<div><div>Thanks to the high precision and flexibility, laser powder bed fusion (LPBF) has hugged in producing key components for the fields of aerospace and biomedicine. However, ensuring the consistent of quality during the manufacturing process remains a headache challenge. Online monitoring the state of melt pool and implementing related closed-loop feedback control is a promising solution to improve quality stability. Especially, the combination of online monitoring and deep learning (DL)-based methods is gaining significant traction. Unfortunately, the ‘black-box’ nature of DL models reduces their reliability of prediction. Additionally, the complex multiphysics-coupled nature of the melt pool often causes transient fluctuations that manifest the inter-layer and intra-layer heterogeneity in monitoring data, which deepens the credibility crisis of DL methods and closed-loop control. In this work, we propose a reliable paradigm for identifying the unstable state of melt pool over inter-layer and intra-layer in LPBF, MSRIM (<strong>M</strong>elt pool <strong>S</strong>tate <strong>R</strong>eliable <strong>I</strong>dentification <strong>M</strong>odel). It outputs both predictions and their uncertainties, enabling control systems to dynamically adjust strategies based on confidence levels. Concretely, we analyze and investigate the heterogeneity of processing data caused by fluctuations of melt pool under different scenarios, along with the uncertainties introduced by such heterogeneity. Then, we quantify and decompose the uncertainties from different sources, and provides a reliable foundation for online control of quality. Furthermore, we develop a custom LPBF melt pool full-processing acoustic emission (AE) monitoring system and created an AE-based dataset including 36 groups of parameters with three melt pool states for verifying our work. Extensive experiments demonstrate that our paradigm achieves the satisfactory and reliable melt pool state identification.</div></div>","PeriodicalId":7172,"journal":{"name":"Additive manufacturing","volume":"109 ","pages":"Article 104887"},"PeriodicalIF":11.1000,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Uncertainty-driven trustworthy identification paradigm for unstable melt pool state based on acoustic emission in LPBF\",\"authors\":\"Jiafeng Tang , Kunpeng Tan , Junlong Tang , Zhibin Zhao , Xingwu Zhang , Xuefeng Chen\",\"doi\":\"10.1016/j.addma.2025.104887\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Thanks to the high precision and flexibility, laser powder bed fusion (LPBF) has hugged in producing key components for the fields of aerospace and biomedicine. However, ensuring the consistent of quality during the manufacturing process remains a headache challenge. Online monitoring the state of melt pool and implementing related closed-loop feedback control is a promising solution to improve quality stability. Especially, the combination of online monitoring and deep learning (DL)-based methods is gaining significant traction. Unfortunately, the ‘black-box’ nature of DL models reduces their reliability of prediction. Additionally, the complex multiphysics-coupled nature of the melt pool often causes transient fluctuations that manifest the inter-layer and intra-layer heterogeneity in monitoring data, which deepens the credibility crisis of DL methods and closed-loop control. In this work, we propose a reliable paradigm for identifying the unstable state of melt pool over inter-layer and intra-layer in LPBF, MSRIM (<strong>M</strong>elt pool <strong>S</strong>tate <strong>R</strong>eliable <strong>I</strong>dentification <strong>M</strong>odel). It outputs both predictions and their uncertainties, enabling control systems to dynamically adjust strategies based on confidence levels. Concretely, we analyze and investigate the heterogeneity of processing data caused by fluctuations of melt pool under different scenarios, along with the uncertainties introduced by such heterogeneity. Then, we quantify and decompose the uncertainties from different sources, and provides a reliable foundation for online control of quality. Furthermore, we develop a custom LPBF melt pool full-processing acoustic emission (AE) monitoring system and created an AE-based dataset including 36 groups of parameters with three melt pool states for verifying our work. Extensive experiments demonstrate that our paradigm achieves the satisfactory and reliable melt pool state identification.</div></div>\",\"PeriodicalId\":7172,\"journal\":{\"name\":\"Additive manufacturing\",\"volume\":\"109 \",\"pages\":\"Article 104887\"},\"PeriodicalIF\":11.1000,\"publicationDate\":\"2025-07-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Additive manufacturing\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2214860425002519\",\"RegionNum\":1,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, MANUFACTURING\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Additive manufacturing","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2214860425002519","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, MANUFACTURING","Score":null,"Total":0}
Uncertainty-driven trustworthy identification paradigm for unstable melt pool state based on acoustic emission in LPBF
Thanks to the high precision and flexibility, laser powder bed fusion (LPBF) has hugged in producing key components for the fields of aerospace and biomedicine. However, ensuring the consistent of quality during the manufacturing process remains a headache challenge. Online monitoring the state of melt pool and implementing related closed-loop feedback control is a promising solution to improve quality stability. Especially, the combination of online monitoring and deep learning (DL)-based methods is gaining significant traction. Unfortunately, the ‘black-box’ nature of DL models reduces their reliability of prediction. Additionally, the complex multiphysics-coupled nature of the melt pool often causes transient fluctuations that manifest the inter-layer and intra-layer heterogeneity in monitoring data, which deepens the credibility crisis of DL methods and closed-loop control. In this work, we propose a reliable paradigm for identifying the unstable state of melt pool over inter-layer and intra-layer in LPBF, MSRIM (Melt pool State Reliable Identification Model). It outputs both predictions and their uncertainties, enabling control systems to dynamically adjust strategies based on confidence levels. Concretely, we analyze and investigate the heterogeneity of processing data caused by fluctuations of melt pool under different scenarios, along with the uncertainties introduced by such heterogeneity. Then, we quantify and decompose the uncertainties from different sources, and provides a reliable foundation for online control of quality. Furthermore, we develop a custom LPBF melt pool full-processing acoustic emission (AE) monitoring system and created an AE-based dataset including 36 groups of parameters with three melt pool states for verifying our work. Extensive experiments demonstrate that our paradigm achieves the satisfactory and reliable melt pool state identification.
期刊介绍:
Additive Manufacturing stands as a peer-reviewed journal dedicated to delivering high-quality research papers and reviews in the field of additive manufacturing, serving both academia and industry leaders. The journal's objective is to recognize the innovative essence of additive manufacturing and its diverse applications, providing a comprehensive overview of current developments and future prospects.
The transformative potential of additive manufacturing technologies in product design and manufacturing is poised to disrupt traditional approaches. In response to this paradigm shift, a distinctive and comprehensive publication outlet was essential. Additive Manufacturing fulfills this need, offering a platform for engineers, materials scientists, and practitioners across academia and various industries to document and share innovations in these evolving technologies.