Study On The Prediction Model For Coke Quality From Wuhan Iron And Steel Corporation

In recent years, people spend more and more attention on coke quality prediction model.Although many prediction models have been established, they all have certain limitation thatcan't be universally apply . In order to establish the coke quality prediction model apply to wisco,this study take the plant of wisco as the background. Through analysis and screening of theexisting coke quality prediction model in domestic and foreign country , selecting reasonablefactors and suitable form of forecasting model, based on analysis and research on laboratoryexperiments on blending coke and production of coking coal, prediction model of coke qualityand production coke quality prediction model in laboratory were respectively predicted usingmultivariate linear regression method,in comparison,the regularity of these two models areentirely consistent, symbol of the corresponding item in these 4 formulas are the same.It showsthat the established prediction models completely reveals the relationship between the cokequality and the coal quality, accroding the production of wisco, the use of V_daf, G, A_d to predictthe cold mechanical strength M₄₀, M₁₀ of coke and the use of V_daf, G, MCI_Y to predict thethermal properties CRI, CSR of coke are feasible.Through the analysis of producted coke and inspect coke quality forecast model ofproducted coke,it found that the forecasted results and actual values matched very well, theaverage error between the forecased results and actual values of coke's cold state strength M10and M40 are 0.07% and 0.24% respectively, the range are 0.19% and 0.45% respectively, theaverage error between the forecased results and actual values of coke's thermal property CRIand CSR are 0.57% and 0.79% respectively, the range are 1.48% and 1.37% respectively, all inallowable error range. Prediction model can fully satisfied the requirement of industrialproduction.