Rolling Force Prediction Of Double-stand Furnace Rolling Mill Based On PSO-BP Neural Network

The mathematical model of the rolling process is the most important part of the modern rolling mill control system,and the mathematical model of the rolling force plays an irreplaceable role in the quality and thickness accuracy of steel products.Nowadays,the quality requirements of steel products are becoming more and more strict.How to improve the quality of rolled products is an important problem to be solved.The accuracy of the rolling force of the double-stand rolling mill affects the quality of the rolled products.The plastic deformation of the metal in the rolling zone is a very complicated process,and there is a strong nonlinear coupling relationship between the parameters of the rolling process and the rolling force.According to the traditional mathematical model formula and research experience,it can not meet the high requirements of the current rolling force accuracy,and can not accurately describe the rolling force variation process of the double stand mill.So this paper uses the rolling pressure mathematical model of rolling mill to combine the neural network and intelligent algorithm to predict the rolling force.Based on the deformation theory of rolling pressure mathematical model,the rolling pressure model of 1725mm double-stand rolling mill in a large rolling mill is taken as the research object,and the basic process parameters of rolling deformation zone are analyzed with Sims and the basic rules of the rolling before and after rolling,etc.,to determine the parameters of the rolling force calculation accuracy of a large number of variables,these parameters are mainly the thickness of rolling,rolling temperature,roll radius,rolling speed;and then rolling force Based on the model,the main variable of rolling force is taken as the input of BP neural network,and the rolling force prediction model with BP network topology of 10-12-1 is established with rolling force as output.The BP network model was trained and tested according to the measured data of Q235 in the large steel mill 1725mm double rack tandem mill.The rolling force of the rolling mill is predicted by the trained neural network,and then the particle swarm optimization algorithm is used to optimize the BP neural network.The prediction results of the two prediction models are carried out.The simulation results show thatThe improved PSO fusion algorithm has the best approximation effect for the rolling force prediction performance,and the PSO optimization BP network algorithm has the same prediction performance.Therefore,this paper finally established a prediction model of rolling force rolling force of double-rack furnace hot rolling mill based on PSO-BP neural network,which effectively improves the prediction accuracy of rolling force of double-rack steckel mill.