Rights statement: ©2022 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Accepted author manuscript, 632 KB, PDF document
Available under license: CC BY-NC: Creative Commons Attribution-NonCommercial 4.0 International License
Final published version
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
}
TY - GEN
T1 - Data-driven Process Parameter Optimisation for Laser Wire Metal Additive Manufacturing
AU - Roberts, Matthew
AU - Xia, Min
AU - Kennedy, Andrew
N1 - Conference code: 27
PY - 2022/10/10
Y1 - 2022/10/10
N2 - Laser Wire Additive manufacturing (LWAM) requires a clear understanding of process parameters and their effects on the geometry and wider material properties of the parts produced to support the production of consistent, repeatable quality parts. Furthermore, its ability to capitalise on using novel alloys depends on efficient characterisation of optimum process parameters. In this work, a method for identifying the range of usable parameters is presented, which produces sufficient data to train Cascade Forward Neural Networks, which are capable of predicting process windows and basic LWAM track geometries for 316L stainless steel. The performance of these networks provides the foundation for further work to identify optimum process parameters and, through transfer learning, may reduce the experimental requirements for the process development of other alloys.
AB - Laser Wire Additive manufacturing (LWAM) requires a clear understanding of process parameters and their effects on the geometry and wider material properties of the parts produced to support the production of consistent, repeatable quality parts. Furthermore, its ability to capitalise on using novel alloys depends on efficient characterisation of optimum process parameters. In this work, a method for identifying the range of usable parameters is presented, which produces sufficient data to train Cascade Forward Neural Networks, which are capable of predicting process windows and basic LWAM track geometries for 316L stainless steel. The performance of these networks provides the foundation for further work to identify optimum process parameters and, through transfer learning, may reduce the experimental requirements for the process development of other alloys.
KW - Additive manufacturing
KW - Laser Wire Metal Additive Manufacturing
KW - Directed Energy Deposition
KW - Machine Learning
U2 - 10.1109/ICAC55051.2022.9911139
DO - 10.1109/ICAC55051.2022.9911139
M3 - Conference contribution/Paper
SN - 9781665498081
SP - 1
EP - 6
BT - Proceedings of the 27th International Conference on Automation & Computing
PB - IEEE
T2 - 27th International Conference on Automation & Computing
Y2 - 1 September 2022 through 3 September 2022
ER -