摘要：The aim of this work was to develop a statisticalmodel which can predict values describingchemical composition of cokes performedin industrial scale. This model wasdeveloped on the basis of data that weretaken from the production system used inthe one of Polish coking plant. Elaboratedequation include quality parameters of initialcoals that form coal blends as well ascontribution of additions such as coke andpetrochemical coke. These equations allowto predict chemical composition of coke, e.g.contributions of: sulphur, ash, phosphorusand chlorine within the coke. A model waselaborated with use of STATISTICA 10 programand it is based on factor and multiplyregression analyses. These analyses werechosen from among few kinds of regressionanalyses. They allowed to develop predictionmodel with the required goodnessof fit between calculated and actual values.Goodness of fit was elaborated with:• residuals analyses,• residues normality and predicted normality• mean absolute error• Pearson correlation confidence
关键词：coke; chemical composition;prediction;mathematical model
