Performance Evaluation of LVS-Based Weld Seam Recognition for Lap Welding Automation in LNG Carrier Cargo Tanks

In the contemporary shipbuilding industry, the construction of Liquefied Natural Gas (LNG) carriers represents a pinnacle of high-value manufacturing that demands extreme precision. The fabrication of membrane-type containment systems, such as Mark III and NO96, relies heavily on the lap-joint welding of thin corrugated plates made of STS 304L or Invar alloy. Recently, the imperative for welding automation has intensified due to the critical shortage of skilled welders and the stringent requirements for gas leakage prevention. However, the application of Laser Vision Sensors (LVS) for seam tracking has been hindered by specular reflection, intense arc interference, and mechanical interference from clamping devices. This study aims to systematically evaluate the seam recognition performance of an LVS-based system under these adverse conditions. A robotic welding system integrated with a coaxial LVS module was constructed to conduct experiments on STS 304L corrugated lap joints. The influence of optical and geometric parameters, including laser power and camera angle, on the quality of acquired laser profiles was analyzed quantitatively. To enhance robustness, the sensor’s logarithmic gain settings were optimized to mitigate interference from TIG arc under varying heat inputs. The results demonstrate that a 200mW line laser at a 50-60° camera angle provides the highest height-difference detection resolution for lap joints. Finally, this research confirms the optical feasibility of LVS in LNG tank fabrication and suggests future directions for hybrid or AI-enhanced sensing solutions.