US 12,465,993 B2
Sensor fusion for welding quality monitoring
Quan Zhou, West Bloomfield, MI (US)
Assigned to Hitachi, Ltd., Tokyo (JP)
Filed by Hitachi, Ltd., Tokyo (JP)
Filed on Jan. 10, 2023, as Appl. No. 18/152,587.
Prior Publication US 2024/0227052 A1, Jul. 11, 2024
Int. Cl. B23K 9/095 (2006.01); B23K 31/12 (2006.01); G05B 19/4155 (2006.01); G06N 20/00 (2019.01)
CPC B23K 9/0953 (2013.01) [B23K 31/125 (2013.01); G05B 19/4155 (2013.01); G06N 20/00 (2019.01); G05B 2219/45104 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
utilizing at least one or more thermal graphic cameras to intake sensor data associated with an arc weld from a robotic welding process, the sensor data comprising thermal imaging data comprising surface geometry, heat/temperature distribution, distribution of high temperature objects, and dynamic change associated with the surface geometry, heat/temperature distribution, and the distribution of the high temperature objects as obtained from the at least one or more thermal graphic cameras;
executing a machine learning model on the sensor data, the machine learning model configured to intake the sensor data and the thermal imaging data as welding process data to output predicted internal parameters of a weld seam of the are weld, predicted surface parameters of the weld seam of the are weld, and a predicted quality of the are weld associated with the predicted internal parameters and the predicted surface parameters; and
controlling the robotic welding process of the arc weld in real-time through modifying parameters of the robotic welding process of the are weld in-real time based on the output predicted quality, the output predicted internal parameters, and the predicted surface parameters.