Digitalization evaluation of welder performance

Abstract

This research studied the natural hand-movement of welder in Gas Metal Arc Welding (GMAW) process. In terms of semi-automatic GMAW process, welder skills could affect the weld bead quality. To monitor torch movement and arc stability of manual welding, the arc signal of current and voltage from the welding machine were utilized to observe on hand movement of welder. The research involved three main stages. First, was simulation of the hand-fluctuation using robot welding, then the Cyclogram technique was constructed by welding current and voltage. The second stage involved with the collection of welding data from the group of experts, intermediate, and beginner welders for comparison of the welding skill in terms of Cyclogram characteristic. The last stage used ellipse fit to generate the quantitative data from Cyclogram for the classification of the welder skills based on the average of ellipse area and the deviation of ellipse area that measures in each second for 20-second period. The results showed the ellipse fit data can differentiate the welder skills based on the statistical analysis. The expert welders produced the mean of ellipse area 5.07 a.u and deviation 1.93 a.u, the intermediate welders got a mean 9.05 a.u and deviation 4.69 a.u, the beginner welders got a mean 18.93 a.u and deviation 8.83 a.u (×104) respectively