Gas metal arc welding process monitoring and quality evaluation using neural networks

Abstract To ensure product quality, it is essential to ensure process quality. Thus, early monitoring and detection of process disturbances in welding production lines are of great significance. The present paper introduces a neural network system for process monitoring and quality evaluation in gas metal arc welding. The system is based only on the measured and statistically processed data for welding voltage and short circuiting time. It is a self-organising feature map Kohonen network which can automatically recognise and classify process disturbances occurring during welding.