Method for predicting coke quality through coking coal nonlinear optimization coal blending

Abstract

The invention provides a method for predicting coke quality through coking coal nonlinear optimization coal blending and provides an important technical guarantee for keeping the stability of coke produced by coke making enterprises and improving the coke quality. The method is characterized in that the indexes of the coke are predicted through a support vector machine technique, a single coal caking property index including two factors, i.e. a gelatinous layer index Y value and a caking index G value, coal blending ratio, a coal and rock index as shown by the accompanying drawing, coking time t and coke oven standard flue temperature T (machine side temperature Tm and coke side temperature Tk) are used as input parameters, then the mechanical strength M40 and M10, the hot performance CRI (Coke Reactivity Index) and CSR (Coke Strength after Reaction) of the coke are used as output parameters, a nonlinear relationship between the input parameters and the output parameters is obtained by training the support vector machine and accordingly the mechanical strength and the hot performance index of the predicted coke are obtained. The method for predicting coke quality through coking coal nonlinear optimization coal blending has the advantages that the influence factors in various aspects in the coke making process are fully considered and the prediction results are enabled to be more scientific and accurate.

Description

technical field [0001] The invention belongs to the technical field of coal coking, and relates to a method for predicting the quality of coke, in particular to a method for predicting the quality of coke by nonlinear optimization of coking coal blending. Background technique [0002] The nature of coking coal is an important factor that restricts the economic benefits of coking enterprises and affects the stable operation of large-scale blast furnace production. Because coking enterprises are faced with many coal mines and variable coal quality, the mechanical strength of coke predicted by large iron and steel complexes through the properties of a single coking coal (M 40 , M 10 ) and thermal performance (CRI, CSR), which have important practical significance for reducing the cost of coking coal blending in coking enterprises. At present, most coking enterprises use the traditional empirical coal blending method to predict the coke quality index, that is, through the sing...

AI Technical Summary

Problems solved by technology

Therefore, although the method reduces the number of input variables, it does not show the effect of coal quality corresponding to each coal type

Images