Nonlinear modeling for 3D weld pool characteristic parameters in GTAW

Modeling and control of the 3D weld pool surface characterized by its width, length, and convexity is an essential capability of the next generation of automated welding machines. In this paper, a nonlinear Hammerstein model is developed for the 3D weld pool surface. A real-time weld pool surface monitoring system is used to measure the specular 3D weld pool surface under a strong arc in gas tungsten arc welding (GTAW). Experiments are designed to produce random changes in the welding current and speed, resulting in fluctuations in the weld pool surface. Static modeling results suggest that the weld pool width and length can be modeled linearly, while the weld pool convexity can be better represented by a nonlinear model. Linear models for the width and length are constructed through a least-squares algorithm- based statistical analysis. A dynamic Hammerstein model is then constructed for the weld pool convexity, and the modeling results are analyzed in detail. The cross couplings between the weld pool characteristic parameters are further modeled using a neuro-fuzzy model. It is found that these nonlinear models can provide a detailed knowledge about the complex correlation between the weld pool surface and welding process inputs.