Self-Learning Fuzzy Neural Networks and Computer Vision for Control of Pulsed GTAW

The objective of this research is to apply intelligent control methodology to improve weld quality. Based on fuzzy logic and artificial neural network theory, a self-learning fuzzy and neural network control scheme has been developed for real-time control of pulsed gas tungsten arc welding (GTAW). Using an industrial TV camera as the sensor, the weld face width of the weld pool, i.e., the feedback signal in the closed loop system, is obtained by computer image processing techniques. The computer vision providing process status information in real-time is an integral part of a self-learning fuzzy neural control system. Such a system enables adaptive altering of welding parameters to compensate for changing environments. The experiments on the control of the pulsed GTAW process show that the scheme presented in this paper can be used to control complicated variables such as encountered in welding processes.