Development of a process model for intelligent control of gas metal arc welding

This paper discusses work in progress on the development of an intelligent control scheme for arc welding. A set of four sensors is used to detect weld bead cooling rate, droplet transfer mode, weld pool and joint location and configuration, and weld defects during welding. A neural network is being developed as the bridge between the multiple sensor set a conventional proportional-integral controller that provides independent control of process variables. This approach is being developed for the gas metal arc welding process. 20 refs., 8 figs.