Application Of Partial Least Squares Regression In BF Iron-making

The iron and steel industry is the foundation of the national industry,which plays an irreplaceable role in developing national economy and Military national defense construction.It is an important index to measure the national economic strength.Blast furnace ironmaking is the first working process of steel industry and usually its energy consumption is the most.With intelligent control we can accurately predict the furnace condition,timely control the furnace,reduce energy consumption and increase the production of pig iron.This article uses the 7#blast furnace of Handan iron and steel company's(volume 2000 m3)1000 groups of actual production data as the research object.Aiming at the limitation of neural network,fuzzy mathematics,chaos and fractal time series:only considering the key parameters or lose valid information.This article adopts partial least-squares regression method(PLSR)to carry out statistics,analysis and modeling of blast furnace data which can improve the accuracy of prediction of furnace temperature.Firstly,the data distribution of the main variables is investigated,and the correlation between variables is investigated,which provides reference and gist for modeling variable selection.Secondly,for the same set of data,the principal component regression and partial least squares regression were used successively.The experimental results are analyzed in detail,and the two ways are compared.It is proved that this method has practical value in blast furnace smelting process.