Research On Optimal Control Of Blast Furnace Coal Injection Based On Multiinformation Fusion

The blast furnace is currently the largest high-temperature closed reaction vessel in the world.Its smelting process has characteristics such as non-linearity,strong coupling,and time delay.Energy consumption accounts for nearly 70% of the total energy consumption of iron and steel enterprises.With the development of intelligent automation,it is imperative to study the optimal control of the blast furnace smelting process.At this stage,replacing coke with coal is one of the most effective ways to achieve energy saving and consumption reduction in blast furnace smelting.However,the control of the timing and amount of coal injection are still artificially controlled by the furnace director,so the operation process is blind and rough.When the coal injection improperly,the furnace conditions will fluctuate violently and major production accidents will occur.In view of the above problems,based on the blast furnace smelting principle,this paper combines the tuyere image information and blast furnace operation data,and uses intelligent control methods to study the optimal control of blast furnace coal injection and conducts the following research work:(1)For the non-contact high-temperature measurement of the blast furnace tuyere swirling area using a sensor,the collected tuyere radiation image is affected by the industrial environment with a lot of noise.This paper uses a wavelet transform and BM3 D filtering algorithm to remove the stripe noise and the Gaussian noise in the tuyere image.At the same time,aiming at the problems of halo and flare in the tuyere image,the watershed algorithm is used to process the tuyere image to extract the target of the tuyere convolution area.Finally,the temperature of the swirling area of the air outlet was calculated via the colorimetric temperature measurement method.(2)On the premise of stable blast furnace conditions and high pulverized coal digestibility,combined with the experience of ironmaking experts,based on the temperature of the tuyere convolution zone and historical operation data of the blast furnace,an expert decision maker for blast furnace coal injection compensation was established to achieve the total effective decision on the amount of coal injectioncompensation.(3)As pulverized coal absorbs heat first and then exotherms when it burns in the gyration zone,if the total compensation amount of the injected coal is added to the blast furnace at one time,which will cause drastic fluctuations in furnace temperature and the furnace conditions deviate from the original equilibrium working point.In response to this problem,this paper uses a predictive control method to make a second decision on the total amount of coal injection compensation.First,the improved PSO-KELM algorithm is used to establish the temperature prediction model of the whirl zone.Next,based on the prediction model,and then through rolling optimization and feedback correction,the predictive control of the coal injection compensation amount is realized.Finally,the total coal injection compensation amount is gradually added to the blast furnace.At the same time,the upper coke amount corresponding to the coal injection compensation amount is replaced by the coal coke replacement ratio,so that the energy flow of the blast furnace is balanced,and the sharp fluctuation of the furnace temperature is avoided.(4)Aiming at the problem that the overall smelting characteristics of the blast furnace have changed after compensation for coal injection,and that there are different blast furnace time lags under different coal injection quantities,this paper adopts the T-S fuzzy control thought.First,the maximum time delay of the blast furnace smelting is determined by combining the experience of ironmaking experts and historical operation data of the blast furnace.Then,divide it artificially.At the same time,the T-S sub-model of the blast furnace coal injection compensation control system is established in different time delay sub-intervals,and the time delay interval is adjusted according to the model approximation accuracy.Finally,the feedback controller is designed based on the principle of parallel distributed compensation,and the stability of the closed loop control system of the blast furnace coal injection is analyzed based on the Lyapunov stability theory and the linear matrix inequality method.Based on the principle of the blast furnace smelting,this paper combines the tuyere image information,blast furnace historical operation data,ironmaking expert experienceand intelligent control methods,a blast furnace coal injection optimization control research scheme is proposed based on the multi-information fusion.Theoretical research and simulation experiments prove that the scheme can guarantee the smooth operation of the blast furnace.This research has important theoretical and practical application value to improve the optimal control level of blast furnace coal injection operation.