Cyber-Physical Optimization of Production Processes Using Cascade AIs: A Robot-Guided MAG Welding Use-Case

Abstract

In last year's article, we proposed a cyber-physical optimization of a robot-guided gas metal arc welding process using two artificial intelligences. These are set up in a cascade to control the welding parameter. This means the first AI performs a rough adjustment of the parameter, while the fine-tuning is done by the second AI. In this paper, we present the results when using this setup. The predicted welding parameter are close to the ideal parameter found experimentally. These findings are already seen with relatively low training data.