An Optimal Coal Injection Decision Model Based On Expert Knowledge And Data

Blast furnace is the largest and most complex single production equipment in industry.The blast furnace smelting process is a dynamic nonlinear system with intermittent and continuous operation modes coexisting,dynamic distribution parameter characteristics,key parameters and operational indicators(hot metal quality,energy consumption and pollution discharge)difficult to detect online.,and it has the characteristics of high temperature sealing,large time delay,multi-phase,strong coupling,non-linearity,time-varying and under-regulation.Blast furnace ironmaking is the most energy-consuming and polluting link in the iron and steel industry chain.It is urgent to transform into a long-lived,efficient,energy-saving,environmentally friendly and efficient intelligent automation production mode.Improving coal injection instead of partial coke is an important means to realize the transformation.However,due to the fluctuation of BF production conditions and the complexity and variability of BF smelting conditions,there are blindness,fuzziness and lag in the timing and quantity of coal injection operation,which makes it difficult to achieve the optimal control objectives of high quality,low consumption and high yield.Therefore,using expert knowledge and process data of blast furnace smelting process to establish control operation optimization model is a hot issue in the field of metallurgy and control,and also a difficult problem to be solved urgently.In view of the above problems,this paper mainly studies the control of lower coal injection in blast furnace smelting process based on expert knowledge and data.The main work is as follows:(1)By reading and studying a large number of documents,we can grasp the mechanism of blast furnace smelting process.According to the characteristics of the mode of operation of the upper minister mechanism and the lower short mechanism of the blast furnace,the optimization control problem of blast furnace operation is equivalent to the optimization control problem of the upper and lower subsystems with constraints by using the hierarchical optimization method.When the blast furnace runs smoothly,the upper burden distribution control is relatively stable,and the effect on the lower temperature control is equivalent to slow disturbance,which simplifies the optimization control of the lower coal injection.Based on the decoupling of the upper part and the lower part,the optimization problem of the lower part of coal injectioncontrol is studied.First,the evaluation model of furnace condition(gas flow distribution and hearth thermal state evaluation model based on furnace temperature prediction)is established,and second,the feedback compensation model of coal injection based on furnace condition evaluation is established.(2)Aiming at the problem of delay in furnace temperature detection and multi-scale characteristics of blast furnace process parameters,a prediction model of furnace temperature based on wavelet multi-scale decomposition Extreme Learning Machine(ELM)is established by using blast furnace process parameters.(3)Reasonable distribution of gas flow and abundant heat in hearth are not only a sign of good furnace condition,but also a sign of high digestibility of pulverized coal.It is also a necessary condition for optimizing coal injection.However,there are many factors affecting blast furnace condition,which are complex and difficult to accurately describe.In this paper,combined with smelting principle,expert knowledge and process parameter information,T-S fuzzy neural network,which combines the complementarity of fuzzy system and neural network,is used to establish a comprehensive evaluation model of furnace condition.(4)Owing to the fluctuation of BF production conditions and the complexity and variability of BF smelting conditions,there are blindness,fuzziness and lag in timing and quantity when operators increase or decrease the amount of pulverized coal injection according to the furnace conditions(furnace temperature).For this reason,the feedback compensation model of coal injection based on the distribution of gas flow and the evaluation of hearth thermal state is established by using the collected parameters data of blast furnace smelting process,which provides guidance for the operation of increasing or decreasing the amount of coal injection by operators.At the same time,the change of coke is calculated by replacement ratio to achieve the purpose of energy saving and coke reduction.This paper takes a large blast furnace in an iron and steel plant as the research object,aiming at achieving high quality,low consumption and high production.From the point of view of feedback compensation control,combining expert experience,process information and intelligent algorithm,the feedback compensation mode of optimum coal injection rate is put forward for the first time,which is suitable for the furnace condition.By using the real-time feedback and rolling correction of the setting value of coal injection,the amount of coalinjection tends to the optimum amount of coal injection adapted to the furnace condition.At the same time,according to the energy flow balance of the blast furnace smelting system,the corresponding coke is replaced by replacement ratio,so that the blast furnace is in a stable state for a long time,which has important scientific significance and broad application prospects for improving the operation optimization control of the blast furnace smelting system.The simulation results show that the control strategy achieves good control effect.