Information Fusion in Pulsed GTAW Using ANFIS and Fuzzy Integral

Welding process is a complex process accompanied by various disturbances, so it is difficult to obtain the welding status by a single sensor. This paper used different sensors to obtain information of the pulsed GTAW (Gas Tungsten Arc Welding) process, and ANFIS (Adaptive Neuro-Fuzzy Inference System) was used to process the information obtained by each sensor, then fuzzy integral method was used to fuse the signal characteristics of each sensor and obtain the prediction results. Experiment results showed that the method could sufficiently use the information obtained by different sensors and obtain more precise prediction results than by using a single sensor alone. This laid the foundation for further controlling the welding quality automatically.