Research And Application Of Centrifugal Granulation Control Of Blast Furnace Slag Based On Neural Network PID

In the iron and steel industry,blast furnace slag is the main by-product of blast furnace ironmaking,which temperature can achieve 1500 ?.The water quenching method commonly used in the industry can not realize the recovery and utilization of its high temperature sensible heat quantity,which also will lead to environmental pollution and waste of resources.Therefore,to solve those problems,the mechanical centrifugal granulation and heat recovery process of blast furnace slag has become the current research hotspot.To realize the recycling of blast furnace slag,the average particle size of granulated slag is required to be less than 2mm,and need the self-adaptive control of centrifugal granulation process.In order to meet the process requirements,in this dissertation,the control method of blast furnace slag centrifugal granulation based on Neural Network PID is proposed.The control parameters are determined on the basis of analyzing the mechanism of centrifugal granulation,and the control method is analyzed by simulation.According to this control algorithm,the automatic control system of centrifugal granulation is developed.The control system is applied to the cold and hot experiments respectively,and the control of centrifugal granulation of blast furnace slag is realized Adaptive control of the process.This thesis is supported by the national key research and development program(2017YFB0603602-03).The main points of this thesis are as follows:(1)By studying the centrifugal granulation mechanism of blast furnace slag granulation,it is clear that the control parameters of slag diameter are slag flow rate and granulation speed.Aiming at the problems that the centrifugal granulation control model cannot be established and the existing slag empirical formula has limitations,this thesis proposes a neural network PID control method combined with the control parameters,and gives the corresponding control strategy.In order to realize the stable control of slag flow,the plug control device was developed and the slag flow was measured and calibrated.(2)In order to verify the effectiveness of the control method,the control method is applied to four typical empirical formula of particle size under different conditions.It can be known from the simulation results,the neural network PID control method can suit for different granulation conditions(3)By analyzing the control requirements of the system,the blast furnace slag centrifugal granulation control system is developed.Taking Mitsubishi PLC as the hardware platform,configuration software and MATLAB as the software platform,the hardware selection and software development of the control system are carried out on the basis of neural network PID algorithm theory.Data transmission is achieved between PLC and MATLAB through OPC communication technology.(4)In order to verify the applicability of the control system,the system is applied to the cold and hot centrifugal granulation experiments.The control system is tested under different working conditions,and the granulation effect of adaptive control and manual control is compared from three aspects of average slag diameter,qualified rate of particle size and slag cotton mass fraction.In the blast furnace slag test results,the qualified rate of particle size increased by 7.2% and the content of slag cotton decreased by 6.7%.In conclusion,the control method of blast furnace slag centrifugal granulation based on Neural Network PID can meet the control requirements of blast furnace slag centrifugal granulation.