| The further advancement of the “double carbon” policy has put forward higher requirements for energy conservation and consumption reduction in the process industry.The cobalt removal purification process is a core step in zinc hydrometallurgy,with numerous and interconnected influencing factors,making the process characterized by multi-variability,nonlinearity,constraints,strong interference,etc.In particular,the production conditions change greatly and key process parameters cannot be measured online.The adjustment of process parameters based on manual operation experience is often subjective,blind and arbitrary,causing problems such as unstable outlet ion concentration of the reactor in the cobalt removal purification process and excessive addition of zinc powder.How to improve the quality of the purified solution while reducing zinc powder consumption is a challenge faced by the current cobalt removal purification process.Combined the current research status with the actual onsite needs of the cobalt removal purification process,firstly the online prediction model of the outlet cobalt ion concentration is focused on in this dissertation.Then,taking model predictive control(MPC)as the main theoretical basis,the optimization control methods applicable to the cobalt removal purification process are studied to ensure that while meeting the process indicators of the enterprise purification process,resource consumption is reduced and production costs are reduced,making it more suitable for complex industrial processes,thereby contributing to the implementation of the “dual carbon” goal.The research contents of this dissertation mainly include the following points:(1)A detailed analysis of the currently used zinc hydrometallurgy processes is provided,and the methods employed in the cobalt removal purification process within these processes are summarized.The antimony salt cobalt removal purification method and the arsenic salt cobalt removal purification method are analyzed and compared,and it is pointed out that these two methods have their own advantages.In view of the abnormality and missing problems in the collected process data due to the influence of various factors,from the perspective of enhancing the generalization ability and improving the prediction accuracy of the model,the detection of abnormal data and the completion of missing data are carried out to provide a solid foundation for mechanism modeling,intelligent modeling,and optimization control.In response to the existing problems and process requirement indicators at the production site,a solution is proposed for predicting and optimizing the outlet cobalt ion concentration of the cobalt removal purification process.(2)The application of dynamic model is of great significance for online prediction of key indicators and process optimization control.Aiming at the problem that the outlet cobalt ion concentration of the antimony salt cobalt removal purification process cannot be detected in time,without changing the actual production process of the zinc smeltery,a hybrid modeling method of combining mechanism modeling and parameter estimation modeling is proposed by considering the complex coupling relationships of the antimony salt cobalt removal purification process.Firstly,the nonlinear overall continuous stirred tank reactor(CSTR)mechanism model of the process is established,and then the deviation between the theoretical value and the actual detected outlet ion concentration is used as the objective function to establish the parameter estimation optimization model.Next,for the problem that the built model cannot be decomposed into linear form,the gradient vector and Hessian matrix of the objective function with respect to the parameter vector are derived,and a solution method based on the steepest descent method and Newton’s method is proposed.Finally,the historical production data of a domestic zinc smeltery are used to accurately invert the model parameters,and the nonlinear dynamic characteristics verification and performance analysis of the model are conducted.The results show that the model has good stability and effectiveness,providing reference for the optimization control of system output.(3)Accurate dynamic modeling of arsenic salt cobalt removal purification process has always been a challenging problem.Aiming at the problem that equating multiple reactors to an overall CSTR model is inconsistent with the actual situation,a dynamic synergistic CSTR(SCSTR)mechanism modeling method for arsenic salt cobalt removal purification process is proposed.Firstly,on the basis of in-depth analysis of the process and reaction mechanism of arsenic salt cobalt removal purification,considering the cascade relationship between the reactors,a SCSTR mechanism model for this process is constructed.Then,for the unknown parameters in the SCSTR model,the idea of Kalman filter is introduced to characterize the unknown parameters as unknown states,and a method for estimating the unknown parameters of the model is proposed using the augmented state equation and unscented Kalman filter(UKF)algorithm.The industrial data simulation results show that the parameter estimation model has high accuracy and the estimated parameters can be used in the SCSTR model.Finally,the dynamic characteristics of the complete SCSTR model are analyzed to verify the influence of different input disturbances on the outlet ion concentration of each reactor,indicating that the SCSTR model has good dynamic performance and provides a momentous basis for the optimization control of the subsequent arsenic salt cobalt removal purification process.(4)Aiming at the problem of how to improve the quality of arsenic salt cobalt removal purification solution while accurately controlling the amount of zinc powder added,a method based on an improved genetic algorithm(GA)backpropagation(BP)neural network combined with distributed nonlinear model predictive control(NMPC)is proposed.Firstly,in view of the problem that a single SCSTR model has difficulty describing the arsenic salt cobalt removal purification process accurately,taking into account the highly nonlinear mapping ability of the data-driven model,a method that organically integrates the SCSTR model with an error compensation model based on GA-BP neural network is proposed(GA-BP-SCSTR)to provide a more accurate online prediction of process indicators.Then,a distributed NMPC architecture is developed using the GA-BP-SCSTR model,control vector parameterization(CVP)technology,and sequential quadratic programming(SQP)algorithm to achieve the coordinated control of the arsenic salt cobalt removal purification process.Finally,the data simulation results of an actual site show that the prediction accuracy of the GA-BPSCSTR model is higher than those of other models,and the proposed predictive control method can maintain the outlet cobalt ion concentrations at the set values while achieving accurate control of the zinc powder addition.This approach can provide good guidance for on-site production.(5)Due to various economic factors,such as energy consumption,operating costs,and changes in customer needs,the operation of the arsenic salt cobalt removal purification process no longer solely considers the issue of meeting production quality standards.How to improve the operation economics of the process has become an urgent problem for the zinc smeltery to solve.To address this problem,firstly a distributed economic model predictive control(EMPC)structure suitable for the arsenic salt cobalt removal purification process is proposed.Then,based on the production reality that zinc powder consumption is the most important economic and technical indicator,an economic performance indicator reflecting zinc powder consumption with terminal constraints is designed to achieve the integration of optimization and control.Next,based on the system dissipative assumption and the design of the rotated cost function,the closed-loop stability of the system is proved.Finally,the simulation experiment shows that the proposed distributed EMPC method can significantly improve the economic performance of arsenic salt cobalt removal purification in the dynamic process by directly optimizing the process economic performance indicator,making a positive contribution to reducing production costs,improving economic benefits,and achieving energy conservation and consumption reduction. |
PDF Full Text Request