A decoupled robotic weld tracking system for multi-layer multi-pass weaving welding of thick T-joint ring stiffeners
Year: 2026
Type: article
Abstract: Thick T-joint ring stiffeners are widely used as critical structural components to enhance the pressure-bearing capacity of ships. However, their large dimensions, substantial thickness, and multi-layer multi-pass weaving welding process pose significant challenges for weld tracking. Currently, its production relies primarily on manual welding, which makes it difficult to ensure consistent welding quality and efficiency. To address this challenge, a 4-axis end-effector is designed and integrated with an industrial robot to construct a robotic weld tracking system. In this system, the robot is responsible solely for following the planned welding path, while the individual axes of the end-effector independently perform tracking and weaving motions. This configuration decouples the various motions involved in the welding process, thereby avoiding the coordinate feedback errors associated with the typical system where the robot performs all motions. Moreover, the structural design of the end-effector decouples the tracking and weaving motions from visual measurement, eliminating their adverse effects on the measurement process. Then, to improve the reliability of long weld tracking, a weld deviation acquisition method including a detection branch and a prediction branch is proposed based on the laser vision sensor. Combined with the above decoupled robotic weld tracking system, this method enables successful weld tracking of thick T-joint ring stiffeners. Experimental results show that this system achieves a mean tracking error of approximately 0.381 mm at the engineering-estimate level, with uniform pass formation and no significant departure at the weld toe, meeting the requirements of practical industrial production.
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $2,950
OpenAlex ID: W7221833145
