Abstract: Gas metal arc welding (GMAW) is widely used in automatic welding especially in robotic welding. The quick response characteristic in robotic welding requires the on line automatic monitoring as well as documentation of welding quality. The electrical signal of welding process is used in order to realize the on line identification of weld quality by processing the electrical parameters. By first analyzing the results of statistical process control, it is found that the method has its limitation. The multivariate statistical process control method is put forward to realize the on line monitoring of weld quality. Experiments have verified its feasibility.

Key words: Multivariate statistical process control, Gas metal arc welding, On line monitoring, Statistical process control, Assembly efficiency, Multi-symmetry features, Part assembly time, Part symmetry