ε-SVM prediction-based rolling force setting calculation method of tandem cold rolling mill
For improving accuracy of the rolling force prediction in the process of the cold rolling, a prediction method based on the support vector machine which optimized by the differential evolution algorithm is proposed. The differential evolutionary algorithm is introduced to the prediction method for optimizing the training parameters of the support vector machine, which can improve the prediction accuracy. The mass production data is used for training the support vector machine, and then the deviation of the rolling force is predicted by the prediction model. The prediction result is applied to modify the rolling force value. Comparing the prediction value with the actual product data, the result shows that this method can efficiently improve prediction accuracy of the tandem cold rolling mill, the relative error decreases from 15% to 6%. Hence, it is an effective method to improve the prediction accuracy of rolling force of the tandem cold rolling mill.
Author's Name: Zhang, S., Jiang, W., Wang, Y.
Volume: Volume 10
Issues: Issue 3
Keywords: Differential evolution, Prediction model, Tandem cold rolling mill, ε-SVM