TY - GEN
T1 - Real-Time Distortion Prediction and Optimization of Weld Sequence Using Artificial Neural Network
AU - Devarai, Jeyaganesh
AU - Ziout, Aiman
AU - Qudeiri, Jaber Abu
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Distortion is one of the predominant challenges concerning the quality and efficiency of a welded joint. It has been shown that distortion is significantly influenced by the weld sequence. In this context, Weld Sequence Optimization (WSO) is ideal for avoiding bottlenecks in the design stage, repairing, and overall capital expenditure in a manufacturing industry. Generally, the weld procedures are determined through experienced welders and in some cases, a plan of experimentation may be necessary. The current research is based on practical testing, computational simulations, and Artificial Neural Networks (ANN). Experiments are carried out using Gas Metal Arc Welding and the Finite Element Model (FEM) has been performed by MSc Simufact Welding software as well as it is validated by high-intensity experiments. The objective of the present research is to create and evaluate a useful method providing real-time predictions of distortion as well as WSO using hot-encoding. The generated optimal sequence from Neural network (NN) models is evaluated by performing the confirmatory test. The findings revealed that the proposed ANN method can significantly predict and optimize weld sequences for reducing distortion.
AB - Distortion is one of the predominant challenges concerning the quality and efficiency of a welded joint. It has been shown that distortion is significantly influenced by the weld sequence. In this context, Weld Sequence Optimization (WSO) is ideal for avoiding bottlenecks in the design stage, repairing, and overall capital expenditure in a manufacturing industry. Generally, the weld procedures are determined through experienced welders and in some cases, a plan of experimentation may be necessary. The current research is based on practical testing, computational simulations, and Artificial Neural Networks (ANN). Experiments are carried out using Gas Metal Arc Welding and the Finite Element Model (FEM) has been performed by MSc Simufact Welding software as well as it is validated by high-intensity experiments. The objective of the present research is to create and evaluate a useful method providing real-time predictions of distortion as well as WSO using hot-encoding. The generated optimal sequence from Neural network (NN) models is evaluated by performing the confirmatory test. The findings revealed that the proposed ANN method can significantly predict and optimize weld sequences for reducing distortion.
KW - Dissimilar Metal Welding
KW - Finite Element Modeling
KW - Gas Metal Arc Welding
KW - MSc Simufact Welding
KW - Sequence Optimization
UR - http://www.scopus.com/inward/record.url?scp=85127800000&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85127800000&partnerID=8YFLogxK
U2 - 10.1109/ICCRD54409.2022.9730142
DO - 10.1109/ICCRD54409.2022.9730142
M3 - Conference contribution
AN - SCOPUS:85127800000
T3 - 2022 IEEE 14th International Conference on Computer Research and Development, ICCRD 2022
SP - 79
EP - 82
BT - 2022 IEEE 14th International Conference on Computer Research and Development, ICCRD 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 14th IEEE International Conference on Computer Research and Development, ICCRD 2022
Y2 - 7 January 2022 through 9 January 2022
ER -