Based on Support Vector Machine of Cold Rolling Force Prediction Research

Abstract

The prediction of the rolling force model is the core of the cold rolling process. When using the traditional mathematical model to calculate the cold rolling force, there is a large calculation error. After analyzing various learning algorithms of artificial intelligence, various models suitable for cold rolling force prediction were studied. A combination of mathematical model and Bayesian LSSVM was proposed. The Bland-Ford-Hill model was used as the main value of pre-calculation of rolling force, and Bayesian LSSVM was used to correct the deviation of the aforementioned calculation of the rolling force model. A large number of actual production data were used for simulation experiments. Through analysis of experimental results, this model can solve the prediction problem under small samples, and has good generalization ability and prediction accuracy.