Research On Mathematical Model Optimization And Control Strategy For Heavy And Medium Plate Controlled Cooling Process After Rolling

The precision of the final temperature in the control cooling process of the heavy and medium plate after rolling has a vitally direct impact on the mechanical properties of steels. It is also the key factor to ensure the shape and quality of the steels. The accuracy of controlling on the final temperature plays an important role in the whole production process of the plate. That illuminates it is of great practical importance to analyze and study the control cooling process after rolling. According to the factual control cooling equipment condition in worksite of an enterprise at home, establishes mathematical model based on the factual cooling equipment condition of the plate plant. The core coefficient of the model is regressed, optimized, off-line put forward, and simulated, in order to improve the control accuracy of the model. The control strategy of the cooling process is mentioned which is based on the above proposed works.The main works and results are as follows:(1) Based on studying the correlative theories and referring to lots of references and materials of heavy and medium plate control cooling process after rolling, the heat transfer theories of plate in the whole cooling process are further researched. The treatment methods for the parameter of physical property and the fixed condition are presented. And according to the boundary and initial condition of thermal conduction inside plate, the solution of finite difference method is researched definitely.(2) The whole control cooling process is divided into three parts, and every part is analyzed seriously. In order to pretty meet the production needs, the heat transfer coefficient is regressed and optimized profoundly. The data, which is from worksite, is collected and analyzed further. The cooling results show that the modified model could improve the final cooling temperature accuracy of plate.(3) Using BP neural network model combined with mathematical model to build the heat transfer coefficient model. The deficiency of traditional BP neural network improved is to decrease the period of BP neural network calculation and improve the precision of the heat transfer coefficient model.(4) The heat transfer coefficient of BP neural network model is off-line trained and tested by the designed simulation programs of7 MATLAB. And used in the laminar cooling system in the same time, the results illuminate that the control precision of the final cooling temperature is improved obviously.(5) According to the worksite condition of an enterprise at home, some feasible online applications are put forward and a corresponding online application model is established. That shows the BP neural network has a wide range of online prospects. The feed forward control strategy used in laminar cooling system is mentioned. The final cooling temperature and the number of open manifolds trend fluctuations could be reduced. That is good to improve the organization, properties and quality of steel products.