Panzhihua Iron And Steel 1220mm Tandem Cold Mill Rolling Schedule Optimization And Model Adaptive Research

The production of cold rolling scrip is a complicated system. There must be passing through many working procedures before raw materials being finished products and it is needed to organize these procedures in an optimized way. Therefore, the connection and the match for all the stands are also an optimization problem. It is an important condition to offer a rational rolling schedule for the high quality, high yield, low energy consumption and environmental protection.Combined with the project of improving the qualities of cold rolling scrip in four-stand tradem cold rolling mills of Pangang Iron &Steel Company, the researchs on rolling schedule optimization and adaptive model have been developed. The calculation of fiction coefficient, deformance resistance, rolling force, rolling torque and rolling power is focused on, and the distribution of thickness of every stand is also done. First, mathematical models are built up in tandem cold rolling mills and are made simple as far as possible in the structure. Therefore, they may have the value of on-line use. Second, according to neual network's characteristic, BP neural network model of friction coefficient is built up. Meanwhile, combined with the adaptive function of BP neural network model, forecast precision of friction coefficient is improved in the process of actual production. Third, forecast precisions of deformance resistance, rolling force, rolling torque, rolling power and trandem rolling tensile force are influenced by the state of rolling process, so forecast procisions of them can not be ensured easily. In order to improve their forecast procisions, adaptive models are done separately. Fourth, on the basis of ensuring the forecast procision of rolling parameters of force and energy, suitable objective functions are built up and the constraint conditions are made necessarily. Then a suitable method (dynamic programming) is chosen to optimize rolling schedules.The findings showed that: predigested mathematical models may meet the needs of on-line use and make the speed of computer higher; Forecast precision of friction coefficient can be improved when BP neural network model is used to on-line forecast; Forecast procisions of rolling parameters of force and eneregy are improved when their adaptive models are carried on; As the objective function on minimum rolling power does optimize the rolling schedules, results are satisfied because optimized thickness of every stand is realized and energy about 1.22% is saved; As the other objective function about the abundant quantity of rolling power does do them, results are that the conserved energy is not so good as the former, but the load coefficient of every stand is close to equally, so the purpose of fully playing the electrical machinery ability has been realized.In the process of trandem cold rolling, establish the rational mathematical models and carry on adaptive model to them; Techniques of artificial intelligence are utilized to forecast the values of rolling parameters and adaptive model are carried on; Optimization techniques are used to optimize rolling schedules. An important role will be played in fully making use of cold rolling mills and improving the cold rolling control precisions.