The current study of weld pool fluid dynamics in arc welding focuses on the numerical model establishment and simulation, and the x-ray combined with particle trace imaging observations, there is no real-time monitor and quantitatively characterize the weld pool flow behavior in welding process for controlling the weld quality. This study develops an innovative structured laser vision-based sensing system for three-dimensional (3D) reconstruction and quantitative analysis of weld pool surface topographies in gas tungsten arc welding (GTAW). Through characterization of dynamic weld pool morphologies, two novel parameters are proposed: the surface convexity variation rate ( R h ) and fluid flow velocity ( R s ), which enable quantitative assessment of weld pool fluid flow characteristics. Complementary two-dimensional(2D) geometrical parameters including weld pool width and length are simultaneously monitored for comprehensive process evaluation. Experimental results demonstrate that the reconstructed 3D pool surface topography combined with the proposed parameters effectively represents the weld pool flow patterns with acceptable accuracy (deviation< 3 %). The convexity gradient R h exhibits strong correlation with flow intensity, where stable variations correspond to optimal bead formation. Quantitative analysis reveals an maximums fluid flow velocity of 70 mm/s, showing strong agreement with computational fluid dynamics simulations. This methodology establishes a new paradigm for in-process monitoring of weld pool dynamics, providing a scientific foundation for intelligent welding quality control.
Zhang et al. (Thu,) studied this question.