Optimization Model Of Energy Consumption For Steelmaking-continuous Casting Process Based On Hybrid Model Of Data-driven And Mechanism

The iron and steel industry is the foundation of our country industrial system,the amount of iron and steel total production for more than a decade has been firmly first in the world,but the iron and steel production energy consumption of per ton steel is far higher than that of developed countries.With the era subject of energy conservation and emission reduction,green environmental protection,how to effectively organize production,reduce the energy consumption and save cost,become a very difficult problem of the steel enterprise at present.This paper is aiming at researching energy consumption of steelmaking-continuous casting production process,through analyzing the actual production process data,using statistical analysis and machine learning methods,suming up and establishing the integrated optimization scheduling model.Then providing a workable solution for steel enterprise.Specific work are as follows:(1)For this article analyzes the existing relevant research results and review,Finding there is a certain gap between the predicted result of oxygen consumption,argon consumption,recycled gas,steam recycled mechanism model and the actual observed value by comparing practical field data statistical results with the analytical results of existing mechanism model.(2)According to the difference in specific model and considering the collinearity condition exists between the multiple variables,using Lasso regression method to selecte the variable automatically and retain the more important variable.As a result,obtaining the linear expression of deviation value and selected variables by simplifying the model.Combining with the corresponding mechanism models comes up with the hybrid models,then the predicted precision has been improved.(3)Using RBF neural network to learning the related process data of mechanism model directly,training network and making a prediction of material consumption and recycled.Compare with the prediction results and accuracy of mixed model and RBF model,then show the result through the system.(4)After analysising the stage and form of energy losses in steelmaking-continuous casting,finds to reduce the temperature drop in the process of transportation and wait in steelmaking-refining and refining-continuous casting can reduce the energy loss effectively.According to the process before and after the transition temperature,steelmaking temperature drop model is set up.At the same time,by analysising the data in steel temperature and drawing speed in casting,finding there is extremely strong linear correlation.with the method of data fitting,linear expression of the pouring temperature and speed are obtained.(5)Energy consumption as objective function to minimize steelmaking-continuous casting process integrated optimization scheduling model is established base on the mixed model,the temperature drop model,the temperature and drawing speed model and workpiece production scheduling model which have been obtained above.Extract more groups of related actual production data and statistic the distribution of the temperature and time,using the result as variable constraint of the model.Using optimization software tools Lingo to solve the model.Verify the validity of the model by using groups of experimental data.