Study On Plant-Wide Optimization Control Based On Interval Number For Hydrometallurgy

As we all know,hydrometallurgy,as one of the techniques for extracting metals from mineral resources,has the advantages of high recovery rate of valuable metals,low environmental protection cost,high product quality,and high process flexibility,and thus it has been widely used.Hydrometallurgy is a complex chemical reaction process with multiple operating conditions,multiple processes and nonlinearities.There is uncertainty in the actual production process,and a single quantitative mechanism model or empirical model can no longer meet the needs of hydrometallurgical process control.Therefore,taking the production process of hydrometallurgy in a gold smelter as the background,based on the mechanism analysis and process analysis of the plant-wide of hydrometallurgy,the plant-wide mechanism model of hydrometallurgy based on interval number is established.The prediction model of hydrometallurgical process is established by using just-in-time learning algorithm and least squares support vector machine,and the overall process optimization and control technology of hydrometallurgy is studied comprehensively.The main content of this thesis is as following:1.Considering that there is uncertainty in the hydrometallurgical process,the expert experience and knowledge is used to establish a qualitative model of the pulping process.Beginning from the basic principles of chemical reaction in the processes of leaching,pressure filtration washing and replacement,based on the relationship between material conservation and energy balance,the static mechanism model of hydrometallurgical plant-wide based on interval number is established.The influence of input variables on production indicators is studied.The simulation validates the validity and rationality of the model established.2.Aiming at the numerous hydrometallurgical processes and complicated working conditions,a strategy for updating the plant-wide of hydrometallurgy is proposed.When the deviation between the prediction value of the process model and the actual value of the industrial process exceeds the set threshold,a kind of just-in-time learning algorithm is used to find a data sample construction learning set with high similarity to the current working point from industrial history data.The predictive model is established by using the least squares support vector machine to update the hydrometallurgical plant-wide model online.The prediction model is applied to the prediction of the first leaching rate,secondary leaching rate and replacement rate of hydrometallurgical plant-wide,and the accuracy and effectiveness of the prediction model are verified by simulation.3.Taking the pulping link as the boundary link,the hydrometallurgical plant-wide optimization problem is divided into several sub-optimization problems.Based on the analysis of the plant-wide of hydrometallurgy,the plant-wide optimization model of hydrometallurgy based on interval number is established,with the maximum comprehensive economic benefit as the optimization goal,sodium cyanide addition quantity,pulping water addition quantity and zinc powder addition quantity as decision variables.Using the two-layer nested optimization algorithm based on improved particle swarm optimization solves the plant-wide optimization model.The optimal operating variables and production indexes of the plant-wide of hydrometallurgy are obtained,which provides guidance for the actual production process,and lays the foundation for the optimization research of the plant-wide of hydrometallurgy with uncertainty.4.Hydrometallurgical process is optimally controlled by a two-layer control structure.In the optimization layer,the optimization algorithm is used to solve the optimization model.The first leaching rate,the secondary leaching rate and the replacement rate are used as the control set value,and the sodium cyanide addition quantity and zinc powder addition quantity are used as the initial value of the control variable.The nonlinear model predictive control method based on standard differential evolution algorithm and improved differential evolution algorithm is used to predict and control the plant-wide of hydrometallurgy.The tracking effect and stability performance of the two control schemes are analyzed.Simulation results demonstrate the superiority of the predictive control scheme based on improved differential evolution algorithm.Finally,the main work of the thesis is summarized,and the further research direction of hydrometallurgical plant-wide modeling and optimization control is discussed and prospected.