期刊
JOURNAL OF MANUFACTURING PROCESSES
卷 71, 期 -, 页码 306-316出版社
ELSEVIER SCI LTD
DOI: 10.1016/j.jmapro.2021.09.033
关键词
Additive manufacturing; Deep learning; Monitoring and control
资金
- National Natural Science Foundation of China [61727802, 61901220]
- China Postdoctoral Science Foundation [2021M691592]
- Postgraduate Research & Practice Innovation Program of Jiangsu Province [KYCX21_0272]
WAAM technology is widely used in metal manufacturing due to its low cost and high efficiency, but issues such as dimensional accuracy, layered morphology, and metallurgical defects limit its further application. This research proposes a method for extracting global information of molten pool and a control method to achieve coordinated monitoring and control of the width and reinforcement in the WAAM process. Experiments show satisfactory control accuracy, providing a feasible way for monitoring and controlling weld shape in the WAAM welding process.
Wire arc additive manufacturing (WAAM) has been used extensively in metal manufacturing and other fields because of its low cost and high efficiency. In the manufacturing process, WAAM technology is often unable to be further applied due to issues such as component dimensional accuracy, layered morphology, and metallurgical defects. To overcome these problems, online monitoring and process control of the welding are necessary. Based on the monitoring of weld width and reinforcement, a regression network for extracting the global information of molten pool is proposed, and an active disturbance rejection control (ADRC) is designed to adjust the welding current. To the best of our knowledge, this is the first time to realize the coordinated monitoring and control of the width and reinforcement of the deposited layer in the WAAM process. Experiments show that the method can obtain satisfactory molten pool width and reinforcement control accuracy, which provides a feasible way for monitoring and control the weld shape in the WAAM welding process.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据