Control Of Hot-roll Pipe Based On Genetic Neural Network

Neural network is one of crucial technologies in the intelligent calculation of 21 century, it has been applied in many fields such as machine learning, optimization control of industry, adaptive control.In recent years ,its application in hot-roll control process has been an important research aspect.Hot-roll process is very complex,its stabilization is relative,but instablization is absolute.This paper is taking the course of hot-roll in the production of seamless steel pipe as object, study the regulation of rotational speed, building mould and optimization method, with the combination strategy of genetic algorithm and artificial neural network, train weight, optimize network right and network structure. Using the course mould, it will get many disturbs for the regulation of rotational speed, and rectify the rotational speed regulating system. So it will make the work of regulating system be near to the best worth , optimize regulation quality,and establish the control system of roll automation that has the characteristic of our country. The main work in the paper includes two following aspects:(1) Have realized regulation of rotational speed in the hot-roll intelligent control system.(2) The result shows that using genetic neural network to forecast and initialize the parameters in the hot-roll control system is feasible.