Intelligent Prediction And Optimization Of Operation Parameters In Aluminum Smelting Process Based On Data Driven

Intelligent manufacturing and intelligent control of process industry has become the focus of industrial research.This article is aimed at the fact that the actual production efficiency of the regenerative aluminum smelting furnace is quite different from the design standard.And the frequent measurement of furnace temperature results in the increase of thermocouple damage,which leads to the increase of production cost.An intelligent prediction and optimization strategy of aluminum smelting furnace temperature based on data driven is proposed.This paper is based on the production process of regenerative aluminum smelting furnace in an aluminum plant in Guangxi.The furnace temperature and operation parameters of aluminum smelting furnace were studied,and some results were obtained.The paper is divided into the following parts:(1)The structure of regenerative aluminum smelting furnace system is introduced.Then,the mechanism of aluminum smelting process is analyzed,and the furnace temperature that can reflect the melting state and the factors that affect the furnace temperature are obtained.Finally,in order to improve the quality of data,the missing values are interpolated by interpolation.According to the three principles,the outliers in the original data are removed,and the whale optimization algorithm is used to optimize the projection pursuit algorithm to analyze the data.The input variables of the soft sensor model of aluminum smelting furnace temperature are selected.(2)However,aluminum smelting process is a complex industrial system with many variables,nonlinear and strong coupling.It is difficult to establish accurate mathematical model based on the traditional methods of physical and chemical mechanism.Therefore,the prediction model of furnace temperature is established by using nuclear limit learning machine.In order to improve the performance of the model,the algorithm is improved.Sine control factor and quantum local search strategy are introduced to get the improved quantum Gray Wolf algorithm.The experimental results show that the algorithm has good performance.Finally,the penalty coefficient and kernel function parameters of the nuclear limit learning machine are optimized,and the prediction model of furnace temperature is established.The results show that the model of furnace temperature prediction established in this paper has high prediction performance.The effective prediction of furnace temperature is realized,which creates conditions for controlling production of aluminum smelting and reducing production cost.(3)On the basis of the above research,the optimal control model of operating parameters of furnace temperature is established.Then,the furnace temperature is divided into 15 temperature sections,and the operation parameters of each section are optimized.Finally,the improved algorithm of quantum gray wolf is used to solve the model.The optimized operating parameters are obtained,namely,the combustion air flow,gas flow,the opening of the combustion air valve,etc.According to the production data,the experimental results show that the optimal control model has good effect.