基于深度学习的钛合金电弧增材制造表面形貌缺陷分类识别与预测

doi:10.11896/cldb.24080130

Abstract
Figure/Table
References
Related Citation (15)

Download:

Full Text ( PDF ) (26434KB)
Export: BibTeX | EndNote (RIS)

Abstract Focused on the defect problems in the forming process of titanium alloy wire arc additive manufacturing (WAAM), This work analyzed the morphology defects such as undercutting, holes, and spheroidization during forming. Firstly, explored the correlation between process parameters and defects in the experiment, obtained the range of process parameters corresponding to each defect, and then deep learning was used to diagnose defects in the forming process of titanium alloy wire arc additive manufacturing. Morphological defects such as undercutting, hole, and spheroidization was classified, identified, and predicted. The search method was used to obtain the best training hyperparameters within a certain range, and the classification performance of four different novel convolutional neural network architectures (CNN) was compared. Image enhancement was used to classify and recognize 4 404 defect images in the forming process for training. The results show that ConvNeXtV2 model has a classification accuracy of 99.06%, with the least training time and parameters, and the best overall performance, the image recognition prediction time is 4.03 ms. It can be used for subsequent feedback control to improve the forming quality of titanium alloy wire arc additive manufacturing.

Key words： titanium alloy wire arc additive manufacturing convolutional neural network morphological defects

Published: 10 July 2025 Online: 2025-07-21

CLC number:

TG444

	Service
	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	MA Ming
	GUO Xinxin
	WEI Zhengying

Cite this article:

MA Ming,GUO Xinxin,WEI Zhengying. Classification and Prediction of Surface Defects in Titanium Alloy Wire Arc Additive Manufacturing(WAAM) Based on Deep Learning[J]. Materials Reports, 2025, 39(13): 24080130-7.

URL:

https://www.mater-rep.com/EN/10.11896/cldb.24080130 OR https://www.mater-rep.com/EN/Y2025/V39/I13/24080130

1 Guo L, He W X, Zhou P, et al.Hot Working Technology, 2020, 49(22), 22(in Chinese).
郭鲤, 何伟霞, 周鹏, 等. 热加工工艺, 2020, 49(22), 22.
2 Feng S R, Ming Z, Bo G S, et al.Ordnance Material Science and Engineering, 2023, 46(4), 101 (in Chinese).
丰生荣, 明珠, 柏关顺, 等. 兵器材料科学与工程, 2023, 46(4), 101.
3 Huang J K, Wu H S, Yu X Q, et al. State of the arc for titanium alloy wire arc additive manufacturing process and microstructure control.Materials Reports, 2023, 37(14), 101(in Chinese).
黄健康, 吴昊盛, 于晓全, 等. 材料导报, 2023, 37(14), 101.
4 Wu S S, Guo C, Liu W M.Modern Manufacturing Technology and Equipment, 2021, 57(3), 204(in Chinese).
吴随松, 郭纯, 刘武猛. 现代制造技术与装备, 2021, 57(3), 204.
5 Ma X, Xiang Z, Li T, et al.International Journal of Minerals, Metallurgy and Materials, 2022, 29(8), 1596.
6 Xiong J T, Geng H B, Lin X, et al.Aerospace Manufacturing Technology, 2015(Z2), 80(in Chinese).
熊江涛, 耿海滨, 林鑫, 等. 航空制造技术, 2015(Z2), 80.
7 Xiao X C, Xin S, Zi R Z, et al.Advances in Materials Science and Engineering, DOI:10.1155/2022/5757484.
8 Akter N S, Song G S.Virtual and Physical Prototyping, DOI:10.1080/17452759.2023.2210553
9 Xia C Y, Pan Z X, Li Y X, et al.The International Journal of Advanced Manufacturing Technology, 2022, 120(1-2), 1.
10 Barua S, Liou F, Newkirk J, et al.Rapid Prototyping Journal, 2014, 20(1), 77.
11 Shen H Y, Du W Z, Sun W J, et al.Applied Sciences, DOI:10.3390/app10010235.
12 Wu J, Huang C, Li Z L, et al.Rapid Prototyping Journal, 2023, 29(5), 910.
13 Li R, Jin M Z, Paquit Vincent C.Materials Design, DOI:10.1016/J.MATDES.2021.109726.
14 Feng Y N. Design andimplementation of surface defect detection system based on wire arc additive manufacturing point cloud. Master’s Thesis, Huazhong University of Science and Technology, China, 2022 (in Chinese).
冯亚宁. 基于电弧增材制造点云的表面缺陷检测系统设计与实现. 硕士学位论文, 华中科技大学, 2022.
15 Gong F J. Study on infrared visual intelligent diagnosis system for defects in CMT additive manufacturing process. Master’s Thesis, Huazhong University of Science and Technology, China, 2020 (in Chinese).
龚福建. CMT增材制造过程缺陷红外视觉智能诊断系统研究. 硕士学位论文, 华中科技大学, 2020.
16 Wan C Z. Research on on-line monitoring of arc welding additive manufacturing system based on multi-sensing. Master’s Thesis, Huazhong University of Science and Technology, China, 2018 (in Chinese).
万柴志. 基于多传感复合的电弧增材制造在线监测系统研究. 硕士学位论文, 华中科技大学, 2018.
17 Ma Chen. Research on Recognition of Morphology Defects and Process Control in TIG-assisted Droplet Deposition Based on Deep Learning. Master’s Tthesis, Xi’an Jiaotong University, China, 2021 (in Chinese).
马琛. 基于深度学习的熔滴复合TIG电弧沉积形貌缺陷识别与工艺控制研究. 硕士学位论文, 西安交通大学, 2021.
18 You X, Cai Y J, Lin Y, et al.Journal of Xiamen University of Technology, 2023, 31(5), 32(in Chinese).
尤鑫, 蔡艺军, 林云, 等. 厦门理工学院学报, 2023, 31(5), 32.
19 Ma Z W, Yuan Y P, Fan P P, et al.Machinery Design and Manufacture, 2023(1), 5(in Chinese).
马占伟, 袁逸萍, 樊盼盼, 等. 机械设计与制造, 2023(1), 5.
20 Wu L J, Qin J Y, Li A H, et al.Journal of Mechanical Transmission, 2022, 46(7), 167(in Chinese).
吴鲁纪, 秦佳音, 李安虎, 等. 机械传动, 2022, 46(7), 167.
21 He K M, Zhang X Y, Ren S Q, et al. In:Conference Record of the 2016 IEEE 29th Conference on Computer Vision and Pattern Recognition. Las Vegas, DOI:10.1109/CVPR.2016.90.