The Preliminary Study Of The Relevance Between Coal Blending And Coke Quality

With the rapid development of the national economy, steel demand is increasing year by year, metallurgical coke used in steel demand is also rapidly increasing, and how quality of coke root stability, and to minimize the cost of coke is coke industry goal.Therefore, blending coking coal through the integration of existing resources to achieve the desired targets of coke, but it is a complex process. Quality of coke with coal, mainly by the nature of, and with the nature of coal is again the nature of the single coal and the ratio of the decision. Therefore, by optimizing with the ratio of coal to coal and coke quality can be controlled to reduce the cost of coking coal.In this paper, reference to a large number of theories and methods of blending home and abroad, based on the analysis of single coal to meet the forecast of coal, that the nature and proportion of single coal with coal, the index predicted a good linear relationship, we can use linear regression, with coal to design reliable prediction model. Prediction of coke quality by linear regression analysis and support vector machine approach, to consider the normal production conditions and process conditions in a wave of cases caused by different factors predicted differences in the predictive analysis of the same test analysis carried out. After calculation, the two factors in the strength of M25 and M10 coke measured and predicted average relative errors were 1.69% and 2.32%, while the five factors M25 and M10 due to the measured value and the predictive value of the average relative error values were 0.69% and 1.50%. Conditions so that there is volatility, the use of this model, five factors can be reliably predicted results. Also proved that the complex relationship blending coking process can be based on statistical theory to correctly describe the support vector machine, in order to further improve the accuracy of prediction of coke quality and provides a new scientific method.In addition, the research group also studied drums with 20 kg test coke into the furnace, the gap between the briquette and the drum is 23 mm, bulk density is 1.2t/m3, the iron test results can predict the industry oven into the coke strength, the prediction model equations are as follows:When the confidence level of 0.95, M25=1.015M25+0.765; M10=1.033M10-2.86;The research group for the first time the use of mathematical simulation test combined with 20kg iron coking coal blending test methods. First use of support vector machine method to predict coke quality, and thus an optimal ratio of model optimization, and then prepared in accordance with the optimal ratio of coal to 20kg drum test validation. In the process, optimizing the blending ratio calculation and coking test method is accurate, can quickly determine the blending ratio, coking enterprises save a lot of material and financial resources, improve enterprise competitiveness.Finally, the research group for the first time the use of mathematical simulation test combined with 20kg iron approach to coking coal blending tests. First, the coke quality prediction model optimization by a number of the best coke ratio, and then use these ratios to 20kg iron test according to test results to determine the final iron ratio. Throughout the process, because the quality of coke and coking model is accurate and reliable test methods, can quickly determine the blending ratio, coking enterprises save a lot of material and financial resources, improve enterprise competitiveness. For the selection of viable economic basis for coal blending program to optimize the blending options, reduce production costs and meet market needs, enhance the market competitiveness of coke. Reasonable guide to blending, blending the ash control to ensure the quality of coke.