Research And Application Of Intelligent Prediction Model For Plan View Pattern Of Plate

Plan view pattern control technology of the plate is a very effective method to make the plate products rectangular,reduce the cutting loss of finished products,and improve the finished product rate.In this paper,the ABAQUS finite element software combined with BP neural network is used to establish the intelligent prediction model of plate shape.Based on the shape control curve parameters and multi-pass rolling process,the plan view pattern of plate can be predicted quickly.Plan view pattern control provides support.The main contents of this paper include:(1)Based on the ABAQUS finite element software,the three-dimensional rolling process of the plate is simulated,the rule of metal flow in the rolling process is analyzed,and the factors such as width,thickness and rolling reduction are studied.According to the simulation results of single-pass and multi-pass of the finite element software,the metal flow of the head and tail of the steel plate is fitted with the curve of four times,and the rolling process of the plate is obtained.Metal flow curve.(2)The BP neural network is used to establish.the intelligent prediction model of metal flow during rolling.The model of the metal flow prediction with the key parameters of the metal flow fitting curve and the inlet thickness,the reduction rate and the width as the input parameters is established,and the parameters of the BP neural network,such as the structure,learning algorithm,hidden layer,input and output parameters,And the results of the finite element simulation as a training and test samples,the BP neural network training and testing.(3)Establishing the plan view pattern prediction model of multi-pass rolling process.Based on the simulation results of BP neural network and ABAQUS finite element rolling,the prediction model of the head and tail shape of the workpiece during the multi-pass rolling process with variable shape control parameters is developed.Based on the plan view pattern parameter curve,In this paper,longitudinal trajectories of different rolling stages are tracked bymultiple tracking points,and the plan view pattern of the finished products is obtained after superposition.(4)Combining with the production process,the plan view pattern prediction program is developed to verify the accuracy and effectiveness of the plan view pattern prediction method.Based on the developed intelligent forecasting method,the plan view pattern of the steel plate is calculated.By comparing the actual production and the model parameters optimization,the plan view pattern evaluation parameters are obtained.It is shown that the plan view pattern prediction algorithm developed in this paper is suitable for engineering application.