Multi-parameter sensing of robotic friction stir welding based on laser circular scanning

Abstract

Sensing is the premise of process control and is of great importance for improving the weld quality in robotic friction stir welding (RFSW). The kinematic characteristics of the 5-axis robot make existing sensing methods unsuitable for the plunge depth and seam deviation. To address this problem, this paper presented a multi-parameter sensing method for the 5-axis RFSW based on laser circular scanning. The laser displacement sensor rotated around the tool forming a series of measured dots to acquire the three-dimensional profile of the workpiece. The measured dots in a circle could be divided into four groups: plane region, welded region, seam region, and disturbed region. Process parameters were extracted using specific algorithms in a circle. The sensing system simultaneously measured the tool tilt angle, plunge depth, and seam deviation online at the maximum frequency of 10 Hz. The absolute measurement accuracy was 0.04°, 0.01 mm, and 0.05 mm for the tool tilt angle, plunge depth, and seam deviation, respectively. And the maximum standard deviation for ten repeated measurements was 0.005°, 0.005 mm, and 0.009 mm, respectively. The feasibility of the sensing system applied to RFSW had been verified through dynamic measurement experiment. This system greatly reduces the number of sensors and facilitates reliable process sensing. Despite being made for 5-axis robots, the sensing system can be used with 6-axis robots and dedicated machines.