Research On Settings Optimization For Reversible Single Stand Cold Rolling Mill

Nowdays, with the development of rolling technology,The metallurgical industry is in high quality and has a high added value, the high technical difficulties of the strip of product demand increased significantly. Each big steel enterprise becomes increasingly fierce competitive in the steel market competition,In order to complete the rolling process of large-scale, quickly running, continuous and automotive development,high precision rolling technology research is imminent.This paper is based on the project of model integration and test system of reversible single stand cold rolling mill of a steel company, research on the optimization for the cold rolling mill.Process control system is the core in cold rolling computer system.That the model calculation, optimal control and cooperate with rolling production is the main task.This paper draw the data flow diagram of level land level3with the process control system by analysised function structure of the level2,at the same time, Confirmed the whole process of optimized model value preset from rolling data receiving,data preparing, preseting calculation,calculating again,back calculation and set data sending to level-1. which make the research clearly,and achieved the sequential design of production preparation modules.In order to obtain a high precision rolling preset model, this paper take the Bland-Ford-Hill model elastic recovery zone and the export thickness calculation into the rolling model, derived the rolling force and flattening radius decoupling of the calculation model, avoid the online iterative process, at same time,estabished rolling model and its son model equation.Using the adaptive control theory of exponential smoothing method to study the rolling model adaptive.Then it studied the back decoupling of the deformation resistance and the friction coefficient and established rolling preset model of the combination of the rolling deformation resistance model and friction coefficient indirect fixed rolling force and adaptive directly fixed rolling force.Using artificial neural network to rolling material parameters preset and adaptive set,because of the BP network’s defuct while it has the slowly convergence speed and easy to the local minimum shortcomings on,this paper designed algorithm of optimizing BP neural network structure and initial right threshold by genetic algorithm. established the BP neural network combined with the adaptive rolling set model, finally improved the rolling setting precision.Finally, this paper used the actual data of steel coils simulating the setting model and analyzed in detail at the end of the proposed research in each section, the results verify this article in design of the feasibility and validity of the method.and finally by establishing the GA-BP algorithm will set accuracy control model of rolling in the3%range, remarkably raised the rolling force set model accuracy.