| Along with the rapid development of national economy, iron and steel industry, as one of the pillar industries, is facing the new challenges as well as great opportunities. In control field, a very important subject is how to apply the new control technology in the iron and steel industry in order to improve quality and enhance the competitive power which finally makes tremendous economic and social benefits.The thesis's background is the tension control of cold-rolling mill. The tension accuracy control is the guarantee of continuous production, so that tension is a key parameter in the rolling process. The research about tension control has been the hot spot and the difficulty of control domain.PSO algorithm is a branch of the intelligence algorithm which is vigorously developing in recent years. Its efficiency on the global optimization arouses researchers'widespread interest. It resolves the contradiction between local and global search, and has stronger and stronger function and compatibility through numerous research and unceasing improvement. Now it is applying in many fields and achieving good results. The thesis researches the PSO algorithm development process, present situation, theory, and improvement direction in detail.The neural network is a widely used intelligence technology. It can approach the random nonlinear function by the random precision through study and training. The thesis particularly analyses the neural network's basic theory and capability, aiming at the disadvantage of the conventional training algorithm, the author proposes a new technique to optimize weights of neural network which uses PSO algorithm.This thesis applies the PSO neural network in PID parameter optimization, to realize the tension intelligent control of cold-rolling mill. Furthermore, it carries on the simulation to the mathematical model of tension control system which establishes in the basic rolling theory. The result indicates that the new technique has achieved the good control effect. |