Prediction and Simulation of cold rolled strip thickness based on PSO-SVM-AdaBoost

Abstract

In order to improve the quality of cold-rolled strip production products, this paper focuses on the problem of thickness prediction accuracy in the production process of cold-rolled strip steel. The rolling data is obtained from the simulation model of the MATLAB. Firstly, this paper comprehensively considers the main influencing factors affecting the cold rolling thickness, secondly, analyses the thickness related parameter variables and combines with the actual production environment, finally a cold rolling thickness prediction model based on PSO-SVM-AdaBoost is proposed. The model is based on the SVM model. The particle swarm algorithm(PSO) optimizes the penalty parameters and kernel parameters of the SVM model. The AdaBoost algorithm integrates weak predictors into strong predictors to improve the prediction performance of the algorithm. In order to illustrate the prediction effect of the model, compared with the prediction results of BP neural network, SVM, PSO-SVM and other models, the experimental results show that the PSO-SVM-AdaBoost model has better prediction performance.