Converter Smelting Static Control Model

Chengde Iron and Steel Corp vanadium extraction steelworks use the Semi steelafter vanadium extraction operation steel-making, because of its converter stove capacityis100t, stove mouth is small, it can not be installed associate sublance operatingequipment, so the control of terminal point always adopting rely on artificial experiencecontrol, this affects production rhythm greatly. To satisfy production needs, change theconverter control present situation of terminal point, raise the carbon of terminal point,is warm to hit the target to lead, establish the control forecast model of static quiet ofterminal point, realize the change that goes to scientific steel-making from experiencesteel-making.According to the site conditions of Hebei Iron and Steel Group Chengde Iron andSteel Corp molten iron and the semi steel smelting, established the "Chengde iron andsteel group100t converter vanadium-steel smelting double static control system". whichis based on neural network prediction model of BP.The research is based static controlmodel of converter smelting, to the basic theory as the basis,which based on the materialbalance and heat balance,based on Visual Basic language as the carrier of BP artificialneural network error model based back propagation network.According to the field of smelting, analysis of factors affecting the end point carboncontent and temperature of the converter, the control variables forecasting model isestablished, separate the carbon content and temperature prediction model, determiningthe input layer has37input variables, intermediate hidden layer has three layers，whichhas30intermediate variables via the BP network. Model training data based on the300furnace samples.The weights and threshold is corrected, and save100furnace non-training learning samples as training models on the basis of the converter, online tainingthrough training model can well adapt to changing production conditions of convertersteelmaking.Which install the converter end point prediction model the smelting production site,to the prediction of end point carbon content and temperature, get a good forecast effect.The semi-steel steelmaking target value of [%C] in the range of±0.02%, the hit rate of80%.[%T] on the target value of±10℃, the hit rate of76.67%. Model meet therequirements of end point carbon content and temperature prediction of converter well.