Research On RBF Algorithm Of Electric Arc Furnace Electrode Adjustment System

In the EAF-steelmaking process,the electrode regulating system is a very important part,position of the electrode and should be real-time and accurate,to adapt to the changes of furnace condition.Current industrial steel-making arc furnace electrode adjustment system generally adopts PID control,although the industry of conventional PID controller can restrain the disturbance of the parameters of the complex system,but due to the complexity of electrode control system of the control object of nonlinear,multivariable,strong coupling,and the furnace condition and parameter time-varying characteristics,such as single PID controller is difficult to efficiently control the electrode,ideal effect is difficult to achieve.Because of the traditional PID parameters setting is difficult to guarantee the real-time online,when furnace condition change is bigger,the existing PID parameters is difficult to satisfy the requirements of furnace condition now,still need to setting of PID parameters,setting time and spend a lot of time.Therefore requires an integrated,intelligent control method to solve the control problem of the complex control system.In this paper,on the basis of consulting a large number of related literature at home and abroad,this paper introduces the electric arc furnace steelmakingequipment,process,both at home and abroad present situation and development prospect of eaf steelmaking,and the characteristics of the electric arc furnace electrode adjustment system and the status quo of the development are summarized.First of all,according to the characteristics of electric arc furnace steelmaking process in this paper,the control strategy analysis and comparison,choose the constant impedance control strategy to control.In the electrical part of the model is deduced on the basis of introducing fuzzy compensation decoupling,by setting appropriate fuzzy rules,the three-phase electrode current impact between small to allow range.Second,a single PID control for electric arc furnace is difficult to meet the conditions of electric arc furnace is complex,in this paper,the Radial Basis Function(RBF)neural network is introduced into the PID control,as a control object identifier,Jacobian information by RBF neural network to control object identification,incremental PID gradient descent algorithm,existing PID parameters setting,establishes RBF-PID electrode regulating system controller,the PID parameters can real-time setting,to suit the requirements of electrode in real time to adjust.Finally,in view of the structure and parameters of RBF neural network model is difficult to determine,respectively,using the particle swarm algorithm and genetic algorithm to optimize RBF neural network parameter values,the optimization of neural network to apply to the readers,the identification error comparison,the results show that the neural network by optimize of PSO,the identification effect is more ideal.