Intelligent laser welding control

Abstract

The development and testing of a laser weld penetration monitoring/control system is described. The system uses an artificial neural network to analyze the air borne acoustic emission from the weld and thus determine weld penetration. This penetration prediction is displayed in real time on a computer screen for operator monitoring and is also used to implement a penetration control. The penetration control uses a fuzzy logic algorithm to calculate welding travel speed changes necessary to maintain penetration at a desired level. The reliability of the system in predicting lap weld penetration was tested and was found to be on the order of 90-95%. Some preliminary results from penetration control tests are shown. They demonstrate that the system remains stable while controlling the weld penetration, but that further work is needed to increase the control accuracy.