A Survey on Intelligentized Technologies for Visual Information Acquirement, modeling and Control of Arc Welding Pool Dynamics

This paper presents a sequential research work on visual information acquirement and intelligent control of arc welding pool dynamics. The visual information acquirement methods are focused in computer vision sensing, image processing and characteristic extraction of the weld pool surface from the single-item pool images by particular algorithms; Based on acquired visual characteristics of weld pool and established neural network, nonlinear and knowledge models for predicting weld pool dynamics during arc welding process, corresponding control methods, such as neural network, self-learning and other composite intelligent control strategies are developed to control welding pool dynamics during pulsed GTAW.