水下
海洋工程
机器人
计算机科学
流量(数学)
系统工程
工程类
环境科学
人工智能
地质学
海洋学
数学
几何学
作者
Qinyun Tang,Ying Du,Zhaojin Liu,Shuo Zhang,Qiang Zhao,Yingxuan Li,Liquan Wang,Tong Cui,Zibing Liu,Gang Wang
摘要
ABSTRACT This paper introduces the design methodology and development of an inspection robot for submerged structures in high‐flow aquatic environments. The robotic system employs dual counter‐rotating vortex suction cups for surface adhesion and achieves omnidirectional mobility through four integrated steering‐drive joints. Its reconfigurable articulated structure enables adaptive conformation to diverse structural geometries. Current regulatory frameworks explicitly mandate non‐destructive testing (NDT) requirements for underwater infrastructure maintenance. Conventional suspended underwater robots face challenges in maintaining stable positional control and consistent surface contact, while existing adhesion‐based systems demonstrate deficiencies in flow‐field adaptability, maneuverability, and transitional capability within complex hydrodynamic conditions. Through systematic exploration of robotic design principles balancing adsorption forces and hydrodynamic resistance under intense flow disturbances, we prototyped and validated the system via experimental campaigns in both dam environments and controlled hydrodynamic test basins. Experimental results demonstrate the robot's capability to execute high‐dexterity maneuvers including in‐situ rotation, comb‐pattern scanning, zigzag trajectories, and intersection curve paths. The system proves effective for inspecting both large‐curvature‐radius surfaces (e.g., dam facades) and small‐diameter tubular structures (e.g., jacket node welds). Field trials confirm that even with simplified control architecture requiring minimal operator intervention, the robot successfully acquires high‐resolution continuous imaging on dam surfaces and obtains valid phased array ultrasonic testing (PAUT) signals for weld inspection, demonstrating detectable sensitivity to artificially induced defects.
科研通智能强力驱动
Strongly Powered by AbleSci AI