Simulation And Research On Nerual Network For Hydraulic Pressure Control Sytem Of 300 Four-High Rolling Mill

With the rapid development of the national economy and the continued progress of science and technology, the quality of steel plates is required by automobile, mechanical manufacturing, electrical and electronics industries. The thickness of plate and strip is one of the most important qualitative indicators. AGC (Automatic Gauge Control) is an indispensable and important component of modern Band plate production.In modern steel rolling production, hydraulic pressing system has already replaced the earlier mechanical and electric pressing system. The positional closed loop control system of hydraulic pressing device combined with various control sections constitutes the hydraulic AGC system. Hydraulic AGC system is a multi-variable, strong coupling, nonlinear and real-time control system. Conventional PID control need to establish a precise mathematical model of controlled object; as to the complex non-linear control system which is difficult dealt with, the introduction of intelligent control technology in the field of rolling has effectively solved such problem and produced a good control effect.Firstly, this paper systematically expounds the basic principles and methods of the thickness control of Band steel. According to specific hydraulic AGC system, considering the nonlinear servo valve load flow and asymmetry of the mass distribution of the load rolling mill system, a dynamic model which controlled by hydraulic AGC positional servo control system is established.Secondly, there is the specific modeling method of the basic element based on the physical structure in software of AMESim. The complete simulation model of the hydraulic AGC electro-hydraulic positional control system is established by using the above method; critical hydraulic components or parameters such as servo valves, hydraulic cylinders, oil bulk modulus and so on effect on dynamic characteristics of hydraulic AGC system is analyzed.Finally, combining with the data processing of the powerful MATLAB software, simulation design of control system is carried out and AMESim / Simulink co-simulation model is built. the main neural network control and PID control combines the design of the BP neural network PID controller, and for the application of BP algorithm easily trapped into local minima, learning the length of the proposed a combination of fuzzy control and neural network control features of the fuzzy RBF neural network PID controller. Yanshan University, 300 experimental rolling mill with hydraulic AGC system objects, the results with the conventional PID control with BP neural network PID control results than the fuzzy RBF neural network PID controller has better control effect.