Optimize Control Of Model Adaptive For Two-Stand Tandem Reversing Steckel Mill

This dissertation focuses on the two-stand tandem reversing steckel mill. Two-stand tandem reversing steckel mill possesses all the rolling mill's characteristics. The tiny tensile force between rolling mill and the canister can be controlled, so it makes steel to roll easily. Furthermore, the rolling mill also has the characteristic of connect rolling mill. It makes the thick steel become thin steel with calefaction, careless roll, extract roll, cooling, etc. Simultaneity, its careless roll and extract roll are processed in a same rolling mill. Rolling must be implemented by different control method to satisfy the requirement of the product quality. And to-and-fro rolling distribution can implement the short flow and energy saving. Two-stand tandem reversing steckel mill can implement the multi-varieties and multi-standards rolling product. This project will require less investment. However, because of the special structures of the two-stand tandem reversing steckel mill, the rolling pressure varies with a great degree in the rolling process, which can result into the thickness and quality of the steel. How to find a control model to match the rolling process perfectly to optimize the process has already became a key problem in the rolling engineering application. Therefore, the research of the process control of rolling mill is of great significance and value.In this paper, the system engineering method is used to analyze the rolling process. The online model, model adaptive learning control method, and adaptive learning optimize method can be found. The model adaptive of rolling control model, careless rolling, and extract rolling can be resolved greatly. The control effect and good performance can be obtained.Concretely, some contents are studied in this paper as follow:Firstly, summary of the research of the rolling control are depicted. Based on the traditional rolling theory, thickness control math model is analyzed and discussed. The model of rolling process control can be analyzed. This dissertation focuses on the relation, correspond of self-learning, adaptive learning control optimize. The online model is founded by least square method and the learning optimize control based on exponential smoothing method is given in this paper.Secondly, distributed control structure of hierarchal multilevel is the control structure of two-stand tandem reversing steckel mill because of the complicate techniques. Control arithmetic given in this dissertation combines the model calculation, self-learning and adaptive technology. Therefore, the control system is a typical adaptive control system. Optimize mechanism of long self-learning and short self-learning based on model adaptive are proposed. Moreover,, model adaptive technology based on model classification and information system classification, the rolling mill strategy optimize method are founded. The fitness of multi-varieties and multi-standards is solved greatly. The control system can optimize the system control parameters automatically.Finally, the operation mode of control system is discussed. After analyzing the on-line simulation, the off-line simulation rolling system is founded with relative practice, which is relative with product and validated by some experiments.The level₂ mathematics model, control system structure and rolling process are researched thorough with the simulation rolling system. Practical application shows that the initial parameters of the rolling model are optimized preferably. The method can save the time and costs.In a word, steel rolling is complicated system engineering. Experiment application shows that the proposed method can receive good control result and improve the real-time character, fitness and coherence of the control model. The system is reliable, stable and higher precision of control. Especially, the methods, which used to optimize the steel enterprise yield process control system, have practical significances and promotional value.