Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM)
Wire arc additive manufacturing (WAAM) is being extensively used in various industrial fields. In WAAM, if a bead is deposited without considering the central angle, its shape may collapse with increasing number of layers. To address this problem, a new method for optimizing the bead geometry using...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-01-01
|
| Series: | Heliyon |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405844023105809 |
| _version_ | 1827367211975573504 |
|---|---|
| author | Dong-Ook Kim Choon-Man Lee Dong-Hyeon Kim |
| author_facet | Dong-Ook Kim Choon-Man Lee Dong-Hyeon Kim |
| author_sort | Dong-Ook Kim |
| collection | DOAJ |
| description | Wire arc additive manufacturing (WAAM) is being extensively used in various industrial fields. In WAAM, if a bead is deposited without considering the central angle, its shape may collapse with increasing number of layers. To address this problem, a new method for optimizing the bead geometry using a support vector machine (SVM) classifier was established in this study. The ranges of the optimal deposition conditions were determined using the SVM classifier and verified by experiments. Geometric data of deposited beads were extracted using a laser profiler, and an SVM binary classifier was used to predict suitable ranges of the deposition conditions. Data were extracted through 20 single-layer basic experiments, classification was performed based on 4°, and the appropriateness of SVM classification was found through 8 single-layer and 3 multi-layer verification experiments.The results showed that the SVM classifier successfully selected the ranges of the optimal deposition conditions. Verification experiments revealed that the results in all cases were appropriately classified based on the boundary of the classification line. Moreover, the SVM classifier was efficient even when a small amount of input data was available. The contribution of this study is that the developed method can help build desired bead geometries in scenarios where deposition is required in the WAAM process, such as re-manufacturing. Thus, this method can be used in real-world industrial applications through further research on the bead shape with multi-layer deposition. |
| first_indexed | 2024-03-08T09:03:14Z |
| format | Article |
| id | doaj.art-329e1ee9d0b74b9c9ac94c3e8bf81c05 |
| institution | Directory Open Access Journal |
| issn | 2405-8440 |
| language | English |
| last_indexed | 2024-03-08T09:03:14Z |
| publishDate | 2024-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Heliyon |
| spelling | doaj.art-329e1ee9d0b74b9c9ac94c3e8bf81c052024-02-01T06:31:37ZengElsevierHeliyon2405-84402024-01-01101e23372Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM)Dong-Ook Kim0Choon-Man Lee1Dong-Hyeon Kim2School of Smart Manufacturing Engineering, Changwon National University, Changwon, Republic of KoreaMechatronics Research Center, Changwon National University, Changwon, Republic of Korea; Corresponding author.Mechatronics Research Center, Changwon National University, Changwon, Republic of Korea; Corresponding author.Wire arc additive manufacturing (WAAM) is being extensively used in various industrial fields. In WAAM, if a bead is deposited without considering the central angle, its shape may collapse with increasing number of layers. To address this problem, a new method for optimizing the bead geometry using a support vector machine (SVM) classifier was established in this study. The ranges of the optimal deposition conditions were determined using the SVM classifier and verified by experiments. Geometric data of deposited beads were extracted using a laser profiler, and an SVM binary classifier was used to predict suitable ranges of the deposition conditions. Data were extracted through 20 single-layer basic experiments, classification was performed based on 4°, and the appropriateness of SVM classification was found through 8 single-layer and 3 multi-layer verification experiments.The results showed that the SVM classifier successfully selected the ranges of the optimal deposition conditions. Verification experiments revealed that the results in all cases were appropriately classified based on the boundary of the classification line. Moreover, the SVM classifier was efficient even when a small amount of input data was available. The contribution of this study is that the developed method can help build desired bead geometries in scenarios where deposition is required in the WAAM process, such as re-manufacturing. Thus, this method can be used in real-world industrial applications through further research on the bead shape with multi-layer deposition.http://www.sciencedirect.com/science/article/pii/S2405844023105809Wire arc additive manufacturing (WAAM)Machine learning (ML)Support vector machineCentral angle |
| spellingShingle | Dong-Ook Kim Choon-Man Lee Dong-Hyeon Kim Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM) Heliyon Wire arc additive manufacturing (WAAM) Machine learning (ML) Support vector machine Central angle |
| title | Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM) |
| title_full | Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM) |
| title_fullStr | Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM) |
| title_full_unstemmed | Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM) |
| title_short | Determining optimal bead central angle by applying machine learning to wire arc additive manufacturing (WAAM) |
| title_sort | determining optimal bead central angle by applying machine learning to wire arc additive manufacturing waam |
| topic | Wire arc additive manufacturing (WAAM) Machine learning (ML) Support vector machine Central angle |
| url | http://www.sciencedirect.com/science/article/pii/S2405844023105809 |
| work_keys_str_mv | AT dongookkim determiningoptimalbeadcentralanglebyapplyingmachinelearningtowirearcadditivemanufacturingwaam AT choonmanlee determiningoptimalbeadcentralanglebyapplyingmachinelearningtowirearcadditivemanufacturingwaam AT donghyeonkim determiningoptimalbeadcentralanglebyapplyingmachinelearningtowirearcadditivemanufacturingwaam |