基于谱线比纹理图的激光能量沉积稀释率实时评估

张志芬; 王杰; 白子健; 张琦; 温广瑞; 林益文

doi:10.7512/j.issn.1001-2303.2026.03.10

智能新核增材未来 | 浏览量 : 0 下载量: 37 CSCD: 0

PDF
导出
分享
收藏
专辑

基于谱线比纹理图的激光能量沉积稀释率实时评估
Real-time Evaluation of Laser Energy Deposition Dilution Rate based on Spectral Line Ratio Texture Maps
2026年56卷第3期页码：73-81
收稿：2026-01-20，

修回：2026-02-24，

纸质出版：2026-03-20
DOI： 10.7512/j.issn.1001-2303.2026.03.10
稿件说明：

移动端阅览

张志芬，王杰，白子健，等.基于谱线比纹理图的激光能量沉积稀释率实时评估［J］.电焊机，2026，56（3）：73-81. DOI： 10.7512/j.issn.1001-2303.2026.03.10.

ZHANG Zhifen, WANG Jie, BAI Zijian, et al.Real-time Evaluation of Laser Energy Deposition Dilution Rate based on Spectral Line Ratio Texture Maps[J].Electric Welding Machine, 2026, 56(3): 73-81. DOI： 10.7512/j.issn.1001-2303.2026.03.10.

摘要

针对激光定向能量沉积（LDED）熔覆层强度实时可靠评估难题，提出了基于双路光电倍增管（PMT）谱线比时间尺度纹理图的激光能量沉积稀释率实时评估方法。利用所开发的双路窄带滤光片-PMT采集系统，实现了对基板合金元素和熔覆粉末合金元素等离子体光谱信号的同步采集，进一步通过多尺度滑动窗口将所选合金元素光谱信号比值转换为二维纹理图，有效保留了光谱信号动态结构与时间相关性；最后提出了融合可变形卷积与SKAttention机制的改进深度学习网络SK-D ResNet稀释率分类模型。验证结果表明，该网络在测试集上实现 98.03% 的分类准确率，较传统卷积网络提升4.77%以上。所提方法为核能等领域重大装备关键构件LDED修复与制造过程熔覆层质量的可靠高精度监测提供了有效解决思路。

Abstract

To address the challenge of real-time and reliable evaluation of the strength of laser-directed energy deposition (LDED) clad layers

this paper proposes a real-time evaluation method for laser energy deposition dilution rate based on the time-scale texture map of the spectral line ratio using a dual-channel photomultiplier tube (PMT). A developed dual-channel narrow-band filter-PMT acquisition system is utilized to achieve synchronous collection of plasma spectral signals from substrate alloy elements and cladding powder alloy elements. Furthermore

by employing multi-scale sliding windows

the one-dimensional spectral line ratio signal is converted into a two-dimensional texture map

effectively preserving the dynamic structure and temporal correlation of the spectral signals. Finally

an improved deep learning network for dilution rate classification

named SK-D ResNet

is proposed

which integrates deformable convolution with the SKAttention mechanism. Verification results show that the network achieves a classification accuracy of 98.03% on the test set

representing an improvement of more than 4.77% compared to traditional convolutional networks. The proposed method provides an effective solution for reliable and high-precision monitoring of LDED repair and manufacturing processes for critical components of major equipment in fields such as nuclear energy.

关键词

Keywords

references

Imran M M ， Che Idris A ， De Silva L C ， et al . Advancements in 3D printing： directed energy deposition techniques， defect analysis， and quality monitoring ［J］. Technologies ， 2024 ， 12 （ 6 ）： 86 .

Chen L ， Bi G ， Yao X ， et al . In-situ process monitoring and adaptive quality enhancement in laser additive manufacturing：A critical review ［EB/OL］. arXiv ， 2024 .

Li X ， Siahpour S ， Lee J ， et al . Deep learning-based intelligent process monitoring of directed energy deposition in additive manufacturing with thermal images ［J］. Procedia Manufacturing ， 2020 ， 48 ： 643 - 649 .

温钟淳，关峰，李明，等 . 工艺调控对Ti-6.5Al-2Zr-1Mo-1V激光熔化沉积层的影响机制研究［J］. 热加工工艺， 2026 ， 55 （ 03 ）： 63 - 70 .

WEN Z C ， GUAN F ， LI M ， et al . Study on Influence Mechanism of Process Modulation on Ti-6.5Al-2Zr-1Mo-1V Laser Melting Deposition Layer ［J］. Hot Working Technology ， 2026 ， 55 （ 03 ）： 63 - 70 .

Mazzarisi M ， Angelastro A ， Campanelli S L ， et al . Monitoring of directed energy deposition laser beam of nickel-based superalloy via high-speed mid-wave infrared coaxial camera ［J］. Journal of Manufacturing and Materials Processing ， 2024 ， 8 （ 6 ）： 294 .

张筱睿，李涛，石博文，等 . 融合注意力机制的ResNet熔覆层稀释率识别方法［J］. 机械设计与制造， 2025 （ 11 ）： 1 - 4 .

ZHANG X R ， LI T ， SHI B W ， et al . ResNet Cladding Layer Dilution Rate Recognition Method Fused with Attention Mechanism ［J］. Machinery Design & Manufacture ， 2025 （ 11 ）： 1 - 4 .

Cai Y ， Wang Y ， Chen H ， et al . Molten pool behaviors and forming characteristics in wire-laser directed energy deposition with beam oscillation ［J］. Journal of Materials Processing Technology ， 2024 ， 326 ： 118326 .

郭文涛 . 基于多传感信号的激光粉末床熔融过程熔池特征原位监测研究［D］. 湖北：武汉纺织大学， 2025 .

GUO W T . In-situ monitoring of molten pool characteristics during laser powder bed fusion based on multi-sensor signals ［D］. Hubei ： Wuhan Textile University ， 2025 .

Staffa A ， Palmieri M ， Morettini G ， et al . Sensitivity Analysis of Integrated Sensors Created Through Additive Manufacturing for Monitoring Components Subject to Dynamic Loads ［J］. Engineering Proceedings ， 2025 ， 85 （ 1 ）： 26 - 26 .

田根，朱甫宏，王文宇，等 . 基于机器学习的传感器监测在金属激光增材制造中的应用［J］. 材料导报， 2025 ， 39 （ 02 ）： 179 - 194 .

TIAN G ， ZHU F H ， WANG W Y ， et al . Application of Sensor Monitoring Based on Machine Learning in Metal Laser Additive Manufacturing ［J］. Materials Reports ， 2025 ， 39 （ 02 ）： 179 - 194 .

Chakrobarty A ， Sultan T ， Davidson A ， et al . In-process ultrasonic monitoring of additive manufacturing：impact of track curvature on spectral responses of periodic builds ［J］. The International Journal of Advanced Manufacturing Technology ， 2025 ， 139 （ 7-8 ）： 1 - 15 .

Nagulin K ， Terentev A ， Vasilev I ， et al . Recycling of used powder materials for additive manufacturing by processing in an inductively coupled plasma ［J］. Powder Technology ， 2025 ， 453 ： 120586 - 120586 .

白子键，张志芬，王杰，等 . 基于光谱物理特征感知网络的DED稀释率监测［J］. 焊接学报， 2024 ， 45 （ 11 ）： 95 - 100 .

BAI Z J ， ZHANG Z F ， WANG J ， et al . Dilution rate monitoring of DED based on a spectral physical feature perception network ［J］. Transactions of The China Welding Institution ， 2024 ， 45 （ 11 ）： 95 - 100 .

Jha A ， M R ， Muvvala G . Laser-material interaction and its significance in the laser-directed energy deposition process ［J］. Journal of Manufacturing Science and Engineering ， 2025 ， 147 （ 11 ）： 111001 .

编辑部网址： http：//www.71dhj.com http://www.71dhj.com