A neural network controls the galvannealing process

High quality galvanized steel strip is a need of today's manufacturers of various products. In particular, in the top quality section steel strips for the automotive, building and consumer goods industries only those steel producers will be successful who are applying state-of-the-art process technologies. For this reason, VOEST-ALPINE Industrieanlagenbau mbH (VAI) and VOEST-ALPINE Stahl Linz developed a new galvannealing control system to optimize this metallurgical process. As the latest improvement of the galvannealing control strategy, a neural network controller has been developed by VAI in Cooperation with the Vienna University of Technology Christian Doppler Laboratory for Intelligent Control Methods for Process Technologies. The paper describes the galvannealing process as far as it is necessary for the understanding of the controller functions, the controller structure and its essential functions. Furthermore, the neural network structure used and its integration into the controller system are explained. A discussion of simulation and practical operating results shows the improvements achieved by using a neural network controller in comparison to the conventional controller.