Reference : Prediction of Cu-Al weld status using convolutional neural network
Scientific congresses, symposiums and conference proceedings : Unpublished conference
Engineering, computing & technology : Mechanical engineering
Physics and Materials Science
http://hdl.handle.net/10993/47688
Prediction of Cu-Al weld status using convolutional neural network
English
Mathivanan, Karthik mailto [University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Engineering (DoE) >]
Plapper, Peter mailto [University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Engineering (DoE) >]
21-Jun-2021
10
No
International
Lasers in Manufacturing (LiM)
21-06-2021 tp 24-06-2021
German Scientific Laser Society (WLT e.V.)
Munich (Virtual)
Germany
[en] Aluminum-copper joints ; weld analysis ; weld type prediction ; Laser welding ; Convolution neural network ; Intermetallic phases ; Image processing ; Energy Dispersive X-ray Spectroscopy (EDS) analysis
[en] Welding copper (Cu) and aluminum (Al) result in brittle intermetallic (IMC) phases, which reduces the joint performance. The key for a strong joint is to maintain an optimum amount of Al and Cu composition in the joint. To implement this without the destruction of the sample is a challenge. For this purpose, high-resolution images of the weld zone are utilized after welding. With the image processing technique, the presence of (Al/Cu) material melted is distinguished. Therefore, the different weld type/status like insufficient melt, optimum melt, and excessive melt is detected from the images.
This paper analyses the weld images and applies the convolutional neural network technique to predict the weld type. The microstructure and Energy Dispersive X-ray Spectroscopy (EDS) analysis of the fusion zone for each weld type are correlated to the weld images.
European Regional Development Fund (FEDER)
Researchers ; Professionals ; Students
http://hdl.handle.net/10993/47688
https://www.wlt.de/lim/

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Limited access
LIM_Proceeding.pdfAuthor postprint1.47 MBRequest a copy

Bookmark and Share SFX Query

All documents in ORBilu are protected by a user license.