TY - GEN
T1 - Using Machine Learning Models to Predict Weld Sequence giving Minimum Distortion
AU - Devaraj, Jeyaganesh
AU - Ziout, Aiman
AU - Qudeiri, Jaber Abu
AU - Baalfaqih, Rashfa
AU - Baalfaqh, Nasmah
AU - Alahbabi, Kanna
AU - Alnaqbi, Maitha
AU - Alhosan, Noura
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - One of the most common issues affecting the performance and reliability of a welded junction is distortion. The welding sequence has been found to have a considerable impact on distortions. In this regard, weld Sequence Optimization (WSO) is useful for preventing these constraints at the early designing phase, and minimizing total capital costs in the manufacturing sectors. Welding processes are usually decided by skilled welders, and in certain circumstances, an experimental strategy may be required. In the present study, realistic experimentation, simulation software, and Artificial Neural Networks (ANN) are used for minimizing the distortion in a complex structure. Experimentation is conducted using Gas Metal Arc Welding for a dissimilar joint from Stainless Steel SS304 with Mild Steel AISI1008, and the Finite Element Model (FEM) was created using MSC Simufact Welding solver and confirmed through a variety of trials. The objectives of this paper is to develop and test a practical strategy for predicting distortion induced during welding process and WSO employing ANN model by hot-encoding. The finding revealed that the distortion is reduced by 87.7 % from the maximum distortion obtained during the welding process.
AB - One of the most common issues affecting the performance and reliability of a welded junction is distortion. The welding sequence has been found to have a considerable impact on distortions. In this regard, weld Sequence Optimization (WSO) is useful for preventing these constraints at the early designing phase, and minimizing total capital costs in the manufacturing sectors. Welding processes are usually decided by skilled welders, and in certain circumstances, an experimental strategy may be required. In the present study, realistic experimentation, simulation software, and Artificial Neural Networks (ANN) are used for minimizing the distortion in a complex structure. Experimentation is conducted using Gas Metal Arc Welding for a dissimilar joint from Stainless Steel SS304 with Mild Steel AISI1008, and the Finite Element Model (FEM) was created using MSC Simufact Welding solver and confirmed through a variety of trials. The objectives of this paper is to develop and test a practical strategy for predicting distortion induced during welding process and WSO employing ANN model by hot-encoding. The finding revealed that the distortion is reduced by 87.7 % from the maximum distortion obtained during the welding process.
KW - ANN Model
KW - Dissimilar Metal Welding
KW - GMAW Process
KW - Hot Encoding
KW - Weld Sequence Optimization
UR - http://www.scopus.com/inward/record.url?scp=85128395769&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85128395769&partnerID=8YFLogxK
U2 - 10.1109/ASET53988.2022.9734845
DO - 10.1109/ASET53988.2022.9734845
M3 - Conference contribution
AN - SCOPUS:85128395769
T3 - 2022 Advances in Science and Engineering Technology International Conferences, ASET 2022
BT - 2022 Advances in Science and Engineering Technology International Conferences, ASET 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2022 Advances in Science and Engineering Technology International Conferences, ASET 2022
Y2 - 21 February 2022 through 24 February 2022
ER -