Development of an Intelligent System for Selection of the Process Variables in Gas Metal Arc Welding Processes

Gas metal arc (GMA) welding is extensively employed in the metal industries for a variety of ferrous and non-ferrous metals because of its potential for increasing the productivity and quality of welding which is controlled by the process parameters. The objective of this paper is to develop an algorithm that enables the determination of process variables for optimised bead geometry for robotic GMA welding. It depends on the inversion of empirical Eq. derived from multiple regression analysis of the relationships between the process variables and the bead dimensions using the least-squares method. The method determines directly those variables which will give the desired set of bead geometry. This avoids the need to iterate by a succession of guesses which are employed in the finite element method (FEM). These results suggest that process variables from experimental equations for robotic GMA welding may be employed to monitor and control the bead geometry in real-time.