Abstract: In view of the key problems such as small match range of welding parameters, strong coupling effect among parameters, unstable welding process and poor weld formation in pulsed MIG welding process, a double-variable decoupling control method is proposed. The weld pool width and wire extension signals are both fed back through vision sensing. On the one hand, weld pool width control is realized by changing double-pulse duty cycle to adjust heat input, on the other hand, wire extension is controlled simultaneously by changing wire feed speed. So the weld formation is improved by control and the stability of welding process is ensured at the same time. A self-tuning fuzzy PID controller is designed and a rapid prototyping control platform is developed for pulsed MIG welding of aluminum alloy by using real-time target environment based on xPC. On this basis, double-variable decoupling control test is carried out by adopting compensation decoupling method and weld beads with uniform pool width and good formation are obtained. The results show that the double-variable decoupling control method can realize the control of pulsed MIG welding process of aluminum alloy and obtain good weld formation on the basis of stable welding process.

Key words: Decoupling control, Pulse metal inert-gas welding of aluminum alloy, Vision sensing, Weld width control