Rack Mill Mathematical Model Adaptive Control

With the development of economy, the requirement on the quality of heat rolling of strip steel is much more strict. The rolling technology which had pursued exceedingly on the quantity in the past is now diverting to emphasize the high quality and high additional value of products.In Kun steel (KISCO), 1725mm production line adopted technology, shape control technical and equipment are advanced. We had made a set of shape control method and optimize model control parameter to improve production quality of 1725mm production line by tracking product process and accumulating data and analyzing the factors affecting shape quality.The different level computer control system is the key for modern steel rolling production. This paper presents the different level computer control system and the modern rolling process. Especially the computer control system of the two-stand mill in KISCO and mathematic model are introduced. At the same time it is advanced to development direction of computer system for rolling process.Based on the research of the Level 2 controlling system of the two-stand mill in KISCO, this paper presents a method to promote the productivity of the mill through the study on the mathematic model and hydraulic automatic gauge control (HAGC). It consists temperature model control, gauge model control and shape model control. We have got effective solutions to solve this problem. In order to improve the prediction precision of rolling load with neural network, the self-adaption of the rolling load model is integrated with the BP neural network for the first time. After the analysis on the on-line data of Kunsteel 1725mm hot-strip plant, the predicting precision of neural networks is improve remarkably, the structure of the neural networks is also simplified. This method affords a new idea for prediction of rolling load by using AI.