Research And Implementation Of Strip Thickness Prediction System Based On Improved Whale Optimizing LSSVM

An important indicator to measure the quality of rolled products in the rolling industry is the thickness of strip steel exports.For many years,the production,consumption,and exports of steel from our country has always ranked first in the world.Therefore,the inaccuracy for the forecast of strip thickness will not only affect the quality of steel,but also produce immeasurable waste of raw materials for the production of steel,At present,the application of intelligent control methods in the field of steel rolling is relatively small,and the accuracy of the strip thickness is not high enough.In addition,there are many factors that affect the thickness of the strip,and the influence of these factors is nonlinear and the degree is different,so intelligent control of the rolling process has become a focus of artificial intelligence research.This paper designs an improved whale optimization LSSVM strip thickness prediction system,because LSSVM will be affected by the penalty factor γ and the kernel function parameter σ,which will affect the prediction accuracy and prediction results.In order to select the best parameters,this article chooses to improve the original whale algorithm from two aspects.On the one hand,it makes the whale less likely to fall into the local optimum to improve its generalization ability.On the other hand,On the other hand,optimize the original coefficients of the algorithm to speed up its convergence speed.use the improved algorithm to find the best penalty factorγ and kernel function parameter σ of LSSVM.Finally,the improved whale optimization LSSVM is applied in the rolling industry to predict the strip thickness.The system is designed and developed five modules,namely registration and login,management information,data processing,prediction model construction and thickness prediction modules.In the data entry and preprocessing module,the mutual information calculation method is used to select the factors that have a greater impact on the strip thickness prediction in the original data set,and the min-max method is used to normalize the influencing factors.In the prediction model construction module,the improved whale algorithm is used to find the optimal parameters of LSSVM,and the data set after the above processing is used to construct the prediction model.The measured strip steel data of a domestic steel group is selected as the experimental data to verify the effectiveness of the improved whale optimization LSSVM proposed in this paper.Through experimental comparison,the data results show that the proposed improved whale optimization LSSVM can better predict the exit thickness for strip.After repeated tests,the improved whale optimization LSSVM strip thickness prediction system designed in this paper can meet the quality requirements of steel products in industrial rolling,and the system meets the expected results in terms of function and performance.