How to Accurately Monitor the Weld Penetration From Dynamic Weld Pool Serial Images Using CNN-LSTM Deep Learning Model?
- Award ID(s):
- 2024614
- PAR ID:
- 10338812
- Date Published:
- Journal Name:
- IEEE Robotics and Automation Letters
- Volume:
- 7
- Issue:
- 3
- ISSN:
- 2377-3774
- Page Range / eLocation ID:
- 6519 to 6525
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
Purpose Ultrasonic additive manufacturing (UAM) is a solid-state joining technology used for three-dimensional printing of metal foilstock. The electrical power input to the ultrasonic welder is a key driver of part quality in UAM, but under the same process parameters, it can vary widely for different build geometries and material combinations because of mechanical compliance in the system. This study aims to model the relationship between UAM weld power and system compliance considering the workpiece (geometry and materials) and the fixture on which the build is fabricated. Design/methodology/approach Linear elastic finite element modeling and experimental modal analysis are used to characterize the system’s mechanical compliance, and linear system dynamics theory is used to understand the relationship between weld power and compliance. In-situ measurements of the weld power are presented for various build stiffnesses to compare model predictions with experiments. Findings Weld power in UAM is found to be largely determined by the mechanical compliance of the build and insensitive to foil material strength. Originality/value This is the first research paper to develop a predictive model relating UAM weld power and the mechanical compliance of the build over a range of foil combinations. This model is used to develop a tool to determine the process settings required to achieve a consistent weld power in builds with different stiffnesses.more » « less
An official website of the United States government

