Calculation Model Of Complex Coke Consumption Ratio Based On Yield Prediction For Sintering Process

Iron ore sintering produces qualified sinter for the blast furnace. The solid fuel (mainly coke), which is the main contributor to the high temperature in sintering process, directly affects the yield and quality of sinter and the energy consumption in sintering process. Decreasing the energy consumption, especially coke consumption by optimizing the sintering process correctly is of great significance to reduce the energy consumption of crude steel and greenhouse gas emission. Efficiently calculating and predicting the coke consumption of the sintering process are the pivotal premise to optimize the sintering process to reduce the coke consumption.Firstly, a reasonable coke consumption indicator, which is complex coke consumption ratio, is defined based on the analysis on the mechanism of coke consumption in sintering process and its calculation mode is developed to measure the energy effective utilization and consumption of fuel in sintering process respectively. Predicting the yield accurately is the pivotal premise to predict the complex coke consumption ratio.Secondly, factors affecting the yield is analysed and critical state parameters’predictive models are developed. The yield of the sinter ore is affected by a variety of factors including production operations, physical and chemical reactions in sintering raw material, and furthermore, sintering production is a strongly nonlinear process, making it is hard to analysis the yield’s influence factors. Factors affecting the yield are determined through the integrated use of the mechanism analysis of the physical and chemical reactions and gray correlation analysis of the production data. The critical state parameters affecting the yield are confirmed and so are their predictive models.Lastly, in order to clearly reflect the variables’impact on the complex coke consumption ratio, a cascade model based on the yield prediction is proposed to describe how the variables affect the complex coke consumption ratio. A BP neural network, whose input variables including outputs of the critical state parameters’predictive models, is developed to predict the yield. The cascade model mainly composed of three levels:the first level composed of complex coke consumption ratio calculate model, the second level composed of yield prediction model and the third level composed of burning through point (BTP) temperature and average vertical sintering speed prediction model. A particle swarm optimization combined with chaotic local search is used to optimize the initial connection weights and translation scaling factors of the BP neural network model to improve the prediction accuracy.The simulation result demonstrates that the modeling methods mentioned herein provides an effective way to predict the coke consumption, which serves as a basis to optimize sintering process and reducing the coke consumption.