Improving RSW nugget diameter prediction method: unleashing the power of multi-fidelity neural networks and transfer learning

doi:10.1007/s40436-024-00503-2

Advances in Manufacturing ›› 2024, Vol. 12 ›› Issue (3): 409-427.doi: 10.1007/s40436-024-00503-2

• •

Improving RSW nugget diameter prediction method: unleashing the power of multi-fidelity neural networks and transfer learning

Zhong-Jie Yue^1,2,3, Qiu-Ren Chen^1,3, Zu-Guo Bao^1,3, Li Huang^1,3, Guo-Bi Tan³, Ze-Hong Hou³, Mu-Shi Li⁴, Shi-Yao Huang³, Hai-Long Zhao³, Jing-Yu Kong³, Jia Wang², Qing Liu³

1. Key Laboratory for Light-weight Materials, Nanjing Tech University, Nanjing, 210009, People's Republic of China;
2. School of Advanced Technology, Xi'an Jiaotong-Liverpool University, Suzhou, 215123, Jiangsu, People's Republic of China;
3. Material Academy, JITRI, Suzhou, 215100, Jiangsu, People's Republic of China;
4. School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, People's Republic of China

收稿日期:2023-10-24 修回日期:2023-11-30 发布日期:2024-09-07
通讯作者: Zu-Guo Bao,E-mail:baozuguo@njtech.edu.cn E-mail:baozuguo@njtech.edu.cn
作者简介:Zhong-Jie Yue is a graduate from the Vehicle Engineering program at the School of Mechanical Engineering, Nanjing University of Science and Technology, currently pursuing a first-year Ph.D. at Xi’an Jiaotong-Liverpool University. His main research interests lie in advanced joining technologies and intelligent manufacturing;
Qiu-Ren Chen is a Distinguished Professor of Nanjing Tech University, and technical expert of Material Academy, JITRI. His main research interests include fatigue durability analysis methods, failure behavior modeling of joints and industrial simulation software development;
Zu-Guo Bao received his Ph.D. in Materials Processing Engineering from Nanjing University of Aeronautics and Astronautics. In 2011-2014, He worked in University of Michigan, Ann Arbor as a visiting scholar. He joined Nanjing Tech University in 2020 and is committed to the R&D of digital design and manufacturing of structural materials, material database and CAE tools;
Li Huang is a Doctor, Vice Director of Materials Data Department at Materials Academy JITRI and Guest Professor at Nanjing Tech University. His research interests include AI algorithms and engineering applications (AI for engineering), integrated computational materials engineering (ICME), digital design of material manufacturing and joining process, and material fatigue and fracture;
Guo-Bi Tan is a process engineer, graduated from Sanjiang College with a bachelor’s degree. Currently employed at the Material Academy, JITRI, working on the Material Big Data Platform;
Ze-Hong Hou graduated from Materials Science and Engineering, currently employed as a materials development engineer at the Material Academy, JITRI. He is responsible for working on the thermal properties and micro-characterization of highstrength steel, cast aluminum, and other metals, as well as composite materials;
Mu-Shi Li is a doctor, graduated from Shanghai Jiao Tong University, focuses on industrial applications of deep learning and machine learning, multi-fidelity surrogate modeling, structural optimization design, industrial big data;
Shi-Yao Huang received a Ph.D. degree from Shanghai Jiao Tong University, Shanghai, China. He is currently a researcher at the Materials Academy of JITRI. His research interests include materials microstr ucture and proper ties prediction;
Hai-Long Zhao pursued his studies and research in Materials Science and Engineering from undergraduate to doctoral level at the University of Science and Technology Beijing. Nowadays he is primarily responsible for the construction and operation of the Industrial Application Materials Big Data Platform, as well as the research, development, and management of various key R&D projects both horizontally and vertically;
Jing-Yu Kong is working at the Yangtze River Delta Advanced Materials Research Institute as a machine learning application engineer, responsible for developing and applying machine learning and deep learning methods to solve complex computational problems in materials science;
Jia Wang is currently an assistant professor in Xi’an JiaotongLiverpool University. She received a Ph.D. degree from the Hong Kong Polytechnic University, M.S. degree from KTH Royal Institute of Technology, B.S. degree in communication engineering from Beijing Jiao Tong University. She had visited the University of Southern California as a visiting scholar. Her research interests span the broadly defined areas of graph mining, reinforcement learning, and explainable AI;
Qing Liu is a recipient of the National Distinguished Youth Fund (1998), a JXZ Distinguished Professor appointed by the Ministry of Education (2006), and the Chief Scientist for two consecutive periods in the 973 Program for Magnesium alloys. He is also the leader of an innovative research group funded by the National Natural Science Foundation and a leading talent in the National "Ten Thousand Talent Program" (2016).
基金资助:
This research has been founded by the Construction Project of the National Natural Science Foundation (Grant No. 52205377), the National Key Research and Development Program (Grant No. 2022YFB4601804), and the Key Basic Research Project of Suzhou (Grant Nos. SJC2022029, SJC2022031).

Improving RSW nugget diameter prediction method: unleashing the power of multi-fidelity neural networks and transfer learning

Zhong-Jie Yue^1,2,3, Qiu-Ren Chen^1,3, Zu-Guo Bao^1,3, Li Huang^1,3, Guo-Bi Tan³, Ze-Hong Hou³, Mu-Shi Li⁴, Shi-Yao Huang³, Hai-Long Zhao³, Jing-Yu Kong³, Jia Wang², Qing Liu³

1. Key Laboratory for Light-weight Materials, Nanjing Tech University, Nanjing, 210009, People's Republic of China;
2. School of Advanced Technology, Xi'an Jiaotong-Liverpool University, Suzhou, 215123, Jiangsu, People's Republic of China;
3. Material Academy, JITRI, Suzhou, 215100, Jiangsu, People's Republic of China;
4. School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, People's Republic of China

Received:2023-10-24 Revised:2023-11-30 Published:2024-09-07
Contact: Zu-Guo Bao,E-mail:baozuguo@njtech.edu.cn E-mail:baozuguo@njtech.edu.cn
Supported by:
This research has been founded by the Construction Project of the National Natural Science Foundation (Grant No. 52205377), the National Key Research and Development Program (Grant No. 2022YFB4601804), and the Key Basic Research Project of Suzhou (Grant Nos. SJC2022029, SJC2022031).

摘要/Abstract

摘要： This research presents an innovative approach to accurately predict the nugget diameter in resistance spot welding (RSW) by leveraging machine learning and transfer learning methods. Initially, low-fidelity (LF) data were obtained through finite element numerical simulation and design of experiments (DOEs) to train the LF machine learning model. Subsequently, high-fidelity (HF) data were collected from RSW process experiments and used to fine-tune the LF model by transfer learning techniques. The accuracy and generalization performance of the models were thoroughly validated. The results demonstrated that combining different fidelity datasets and employing transfer learning could significantly improve the prediction accuracy while minimize the costs associated with experimental trials, and provide an effective and valuable method for predicting critical process parameters in RSW.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00503-2

关键词: Resistance spot welding (RSW), Nugget diameter prediction, Multi-fidelity neural networks, Transfer learning

Abstract: This research presents an innovative approach to accurately predict the nugget diameter in resistance spot welding (RSW) by leveraging machine learning and transfer learning methods. Initially, low-fidelity (LF) data were obtained through finite element numerical simulation and design of experiments (DOEs) to train the LF machine learning model. Subsequently, high-fidelity (HF) data were collected from RSW process experiments and used to fine-tune the LF model by transfer learning techniques. The accuracy and generalization performance of the models were thoroughly validated. The results demonstrated that combining different fidelity datasets and employing transfer learning could significantly improve the prediction accuracy while minimize the costs associated with experimental trials, and provide an effective and valuable method for predicting critical process parameters in RSW.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00503-2

Key words: Resistance spot welding (RSW), Nugget diameter prediction, Multi-fidelity neural networks, Transfer learning

Zhong-Jie Yue, Qiu-Ren Chen, Zu-Guo Bao, Li Huang, Guo-Bi Tan, Ze-Hong Hou, Mu-Shi Li, Shi-Yao Huang, Hai-Long Zhao, Jing-Yu Kong, Jia Wang, Qing Liu. Improving RSW nugget diameter prediction method: unleashing the power of multi-fidelity neural networks and transfer learning[J]. Advances in Manufacturing, 2024, 12(3): 409-427.

参考文献

1. Makoto K, Yoshimasa F, Kazuhiro O et al (1996) Estimation of tensile shear strength of spot welded joint of steel sheets. Q J Jpn Weld Soc 14(4):754-761
2. Oikawa H, Murayama G, Hiwatashi S et al (2007) Resistance spot weldability of high strength steel sheets for automobiles and the quality assurance of joints. Weld World 51(3):7-18
3. Pouranvari M, Asgari HR, Mosavizadch SM et al (2007) Effect of weld nugget size on overload failure mode of resistance spot welds. Sci Technol Weld Joi 12(3):217-225
4. Sun DQ, Lang B, Sun DX et al (2007) Microstructures and mechanical properties of resistance spot welded magnesium alloy joints. Mater Sci Eng A 460:494-498
5. Sim J, Kim KY (2018) Hybrid nugget diameter prediction for resistance spot welding. Procedia Manuf 17:395-402
6. Birada AK, Dabade BM (2020) Optimization of resistance spot welding process parameters in dissimilar joint of MS and ASS 304 sheets. Mater Today Proc 26(2):1284-1288
7. Luo Y, Li CT, Xu HB (2009) Regression modeling and process analysis of resistance spot welding on galvanized steel sheet. Mater Des 30(7):2547-2555
8. Pandya KS, Grolleau V, Roth CC et al (2020) Fracture response of resistance spot welded dual phase steel sheets:experiments and modeling. Int J Mech Sci 187:105869. https://doi.org/10.1016/j.ijmecsci.2020.105869
9. Kumar A, Ghosh GK, Panda S et al (2020) Numerical simulation of weld nugget in resistance spot welding process. Mater Today Proc 27:2958-2963
10. Andersson O, Melander A (2015) Prediction and verification of resistance spot welding results of ultra-high strength steels through FE simulations. Model Numer Simul Mater Sci 5(1):26-37
11. Eisazadeh H, Hamedi M, Halvaee A (2010) New parametric study of nugget size in resistance spot welding process using finite element method. Mater Des 31(1):149-157
12. Hussein HK, Shareef IR, Zayer IA (2019) Prediction of spot welding parameters using fuzzy logic controlling. Eastern-Eur J Enterp Technol 5(2):57-64
13. Thongchai A, Kawin S, Phisut A et al (2014) Resistance spot welding optimization based on artificial neural network. Int J Manuf Eng 2014:154784. https://doi.org/10.1155/2014/154784
14. Mallaradhya HM, Kumar MV, Chandra MSV (2022) Optimization of parameters and prediction of response values using regression and ANN model in resistance spot welding of 17-4 precipitation hardened stainless steel. J Adv Manuf Syst 21(2):275-291
15. Tan Y, Fang P, Zhang Y et al (1999) Evaluating nugget sizes of spot welds by using artificial neural network. In:Proceedings of the 6th international conference on computational intelligence, theory and applications:fuzzy days, 25-28 May, Springer, Berlin
16. Panza L, Bruno G, Antal G et al (2024) Machine learning tool for the prediction of electrode wear effect on the quality of resistance spot welds. Int J Interact Des Manuf. https://doi.org/10.1007/s12008-023-01733-7
17. Kitano H, Nakamura T (2018) Combined artificial neural network and least squares method for exploring relationships between welding conditions and weld characteristics. Weld Lett. https://doi.org/10.2207/qjjws.36.5WL
18. Kitano H, Nakamura T (2019) Automatic derivation of empirical formulas for characteristics of weld joints using machine learning based technique. J Jpn Weld Soc 88(7):532-535
19. Kitano H (2019) Numeric law discovery and knowledge extraction from welding phenomena using machine learning technique. Mater Jpn 58(8):449-452
20. Kitano H, Sato A, Iyota M et al (2021) Investigation of relationship between resistance spot welding condition and nugget shape by utilizing machine learning based technique. Weld Int 33(4/6):223-230
21. Gavidel SR, Jeremy L (2019) Performance analysis and comparison of machine learning algorithms for predicting nugget width of resistance spot welding joints. Int J Adv Manuf Technol 105:3779-3796
22. Li M, Liu Z, Huang L et al (2022) Automatic identification framework of the geometric parameters on self-piercing riveting cross-section using deep learning. J Manuf Process 83:427-437
23. Zhang Z, Wen G, Chen S (2019) Weld image deep learning-based on-line defects detection using convolutional neural networks for Al alloy in robotic arc welding. J Manuf Process 45:208-216
24. Dai W, Li D, Tang D et al (2021) Deep learning assisted vision inspection of resistance spot welds. J Manuf Process 62(8):262-274
25. Park C, Haftka RT, Kim NH (2017) Remarks on multi-fidelity surrogates. Struct Multidiscip Optim 55:1029-1050
26. Zhang X, Xie F, Ji T et al (2021) Multi-fidelity deep neural network surrogate model for aerodynamic shape optimization. Comput Method Appl Mech Eng 373:113485. https://doi.org/10.1016/j.cma.2020.113485
27. Zhang Y, Kim NH, Park C et al (2017) Multi-fidelity surrogate based on single linear regression. AIAA J 56(12):1-9
28. Li M, Liu Z, Huang L et al (2023) Multi-fidelity data-driven optimization design framework for self-piercing riveting process parameters. J Manuf Process 99:812-824

Improving RSW nugget diameter prediction method: unleashing the power of multi-fidelity neural networks and transfer learning

Improving RSW nugget diameter prediction method: unleashing the power of multi-fidelity neural networks and transfer learning

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics

本文评价