저온용사된 알루미늄-스테인리스강 이종 레이저 용접부의 특성 평가 및 딥러닝 기반 용입 깊이 분류Characterization of Cold-Sprayed Aluminum-Stainless Steel Dissimilar Laser Welds and Deep Learning-Based Weld Penetration Classification
- Other Titles
- Characterization of Cold-Sprayed Aluminum-Stainless Steel Dissimilar Laser Welds and Deep Learning-Based Weld Penetration Classification
- Authors
- 조영길; 이승환; 최돈현; 강민정
- Issue Date
- Apr-2025
- Publisher
- 대한용접접합학회
- Keywords
- Al/Fe dissimilar welding; Cold spray; Laser welding; Intermetallic compound; Corrosion test; CNN based deep learning
- Citation
- 대한용접접합학회지, v.43, no.2, pp 194 - 204
- Pages
- 11
- Indexed
- KCI
- Journal Title
- 대한용접접합학회지
- Volume
- 43
- Number
- 2
- Start Page
- 194
- End Page
- 204
- URI
- https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/207315
- DOI
- 10.5781/JWJ.2025.43.2.9
- ISSN
- 2466-2232
2466-2100
- Abstract
- Laser welding offers advantages like high speed, narrow seams, and reduced heat-affected zones, but is limited when joining dissimilar materials such as aluminum (Al) and low-carbon steel (Fe) due to differences in physical properties and the formation of brittle intermetallic compounds (IMCs), including FeAl2 and Fe2Al5. To address this, cold spray technology propels Fe powder at high velocity to create mechanical bonding, suppress IMC formation, and enhance interface stability. In this study, laser welding was applied to overlapped joints of stainless steel, a cold-sprayed Fe layer, and aluminum. Mechanical and microstructural properties were evaluated under varying welding parameters and corrosive environments. Additionally, a CNN-based model using thermal and molten pool images from CMOS and IR cameras was developed to classify weld penetration states. The findings confirm cold spray’s effectiveness as an interlayer method and show that AI enables process control over weld penetration.
- Files in This Item
-
Go to Link
- Appears in
Collections - 서울 공과대학 > 서울 기계공학부 > 1. Journal Articles

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.