Research On Temperature Prediction Of Blast Furnace Based On Data And Tuyere Images

The blast furnace iron-making is a very complex industrial system with the followingcomprehensive complexity: Large time delay, field coupling, dynamic distribution parameter,intermittent and continuous operation mode coexist, strong nonlinearity, multivariable coupling andmixed variable type, etc. The operations of the complex characteristic changes over productionconditions, the raw materialcomposition, operation condition, equipment and other uncertain factors,described in precise mathematical model.The temperature prediction is of great significance in the process of blast furnace iron-making.High temperature can make the coke rate increases, higher production costs, and reduce the usefulservice life ofblast furnace; Temperature is too low will make furnace heat ofreaction is insufficient,pig iron production is reduced, and even lead to accident in blast furnace. Process parameters affectthe furnace temperature is very much, many factors cannot predict in advance and are oftenoperating conditions have changed, and wait until the result is shown by the operating personnelwhen found, began to take remedial measures if at this time, had formed the furnace conditionfluctuation. How to help identify factors that influence the furnace condition fluctuation and takecorresponding measures into the blast furnace operatorshave to face a problem.The working conditions of blast furnace tuyere of blast furnace smelting process plays anextremely important role. Operators can through the observation of the tuyere bright degree tounderstand the status of the hearth heat, tuyere peephole CCD radiation image grayscale changereflects the size of the tuyere radiation intensity, can be more accurate characterization of thethermal state of tuyere cyclotron. Existing research the following problems： on the one hand, dueto the complexity of the blast furnace production conditions and the furnace condition fluctuation,the existing prediction models in raw materials, production and operating condition changes canpredict, need according to production condition of a particular blast furnace and the matchingfurnace temperature prediction model is established. On the other hand, the blast furnace is basically a "black box", only can be directly observed from the tuyere peephole inside the blast furnace, theoperator attention as a important basis of furnace temperature. But previous studies rarely considerthe effect ofthe factorsonthe furnace temperature prediction.This article first introduced in the context of energy conservation and emissions reductionfurnace temperature prediction and research progress both at home and abroad, the importance of,some important parameters and tuyere of blast furnace ironmaking technology after the relationshipbetween image and furnace temperature prediction has carried on the detailed description, theresearch work of this paper is focused on. Extraction of tuyere image first, use the digital imageprocessing method to calculate average outlet image grey value, the furnace temperature predictionmodel was established based on the least square method; Then blast furnace data for parameterselection and data preprocessing, furnace temperature prediction model based on the data;Comprehensive utilization of the final image data with the blast furnace tuyere, the establishment ofBP neural network model for molten iron content of [Si] and to predict the molten iron temperature,through the contrast to find the optimal furnace temperature prediction method. The simulationresults prove that this method can predict the furnace temperature, models to predict shooting ishigher than only using or blast furnace tuyere image data to predict. The research for the blastfurnace operatorsto provide a more reliable furnace condition information, to ensure high efficiencyand high yield of blast furnace smelting and provide reference for saving energy and reducingconsumption.