Experimental Calibration of an Automated GMAW Model

This paper describes the calibration of a dynamic model of the gas-metal arc welding process (GMAW). First, a fifth-order, nonlinear dynamic model of the GMAW process is presented. This model is parameterized by various physical constants as well as five empirical constants. Next, we describe experimental data (steady-state current and arc voltage) obtained from an automated GMAW system over a range of operational conditions defined by variations in open-circuit voltage and wire feed speed. Analysis of the data suggests loop pairings to use in a multi-loop control strategy. Then, using an optimization strategy, the "best" value of each parameter is determined, where best is defined relative to the error between the actual GMAW current and that predicted by the model, averaged over all the data The best value of each parameter is also determined for each operating point. Contour plots of the model error show that the empirical "constants" are a function of the operating conditions of the system.