Based On Neural Network Model Of Hot-rolling

With the development of steel rolling technology, users are strict in quality of the product especially in strip thickness and accuracy. Because the strip thickness accuracy depends primarily on the rolling force model setting accuracy, so improvement in the setting accuracy of rolling force model has become the key to solving the problem of the thickness accuracy of strip steel. Based on the neural network research, the passage makes a comparison between the advantages and disadvantages of neural network, RBF network to predict the rolling force change. Rooted on the statistical analysis from a large-scale product, we have experimented a new rolling force neural network prediction model. The new model of the error prediction compared has generally reduced by more than 50 percent, compared with the traditional rolling force calculation model for error prediction.