Research And Design Of Iron Ore Flotation Prediction And Process Decision System Based On Mineral Genes

Flotation is an important process in the mineral separation process,and it is also the focus of research in the field of mineral processing.The flotation process,the reagent system and its dosage have a significant impact on the flotation results.The genetic characteristics of minerals are closely related to the selectability of iron ore,which will affect the selection of iron ore flotation process and the determination of the reagent system and dosage.However,the research on mineral genetic characteristics is insufficient,and the research and development of mineral processing technology and engineering design Insufficient integration of modern information technology.Therefore,to carry out relevant research on the genetic characteristics of minerals,use the beneficiation test data,process mineralogy research data,production data,etc.to establish an iron ore flotation process decision support system,which will help guide the selection of the flotation process and the reagent system,and shorten the development cycle of flotation technology is of great significance to realize the high efficiency and energy saving of iron ore flotation process.This thesis takes mineral gene flotation as the research object,and uses a combination of mechanism and theoretical analysis,experimental research,computer simulation,and data statistical analysis to complete the following tasks:A prediction model for the dosage of iron ore flotation reagents is established,based on experimental design methods.The pharmacy system was optimized,and the mineral gene flotation database and process knowledge base were constructed.Based on the rough set production rule representation,the rule engine reasoning technology was used to establish the iron ore flotation process decision model.The main research results of this thesis are as follows:(1)Aiming at the prediction of flotation reagent dosage for small samples,this thesis proposes a flotation reagent dosage prediction method based on Grey Wolf Optimizer(GWO)combined with Support Vector Regression(SVR)model.The prediction accuracy of the SVR model is closely related to the model parameters.The parameter determination usually uses cross-validation to try to adjust the parameters,which is easy to fall into the local optimum.The GWO algorithm has an adaptively adjusted convergence factor and an information feedback mechanism,and has a great advantage in global optimization.In this thesis,GWO is used to optimize the penalty coefficient and kernel function related parameters in the SVR algorithm.Through comparative experiments on the prediction results and production data,the prediction accuracy reached 90%.(2)The flotation results of iron ore are affected by many factors such as the nature of the original ore,the flotation reagent system,and the dosage of reagents.It is difficult for singlefactor flotation experiments to reveal the relationship between the factors,and it is also difficult to determine the best combination of reagent systems.In response to this problem,this thesis proposes an optimization model of flotation reagent system based on Orthogonal experimental design(Orthogonal experimental design)combined with Response Surface Methodology(RSM).Taking iron grade as the evaluation index,a five-factor three-level flotation test was designed.Through simulation analysis,it is verified that the optimized flotation reagent system is consistent with the laboratory results.(3)Research on knowledge representation,reasoning mechanism and decision-making model establishment method of iron ore flotation process.Knowledge in the field of mineral processing has the characteristics of complex data sources,irregular data,unstructured and relatively scattered,etc.This thesis combines the genetic characteristics of ore-related structural structures and mineral material composition to address the problem of unclear edge rules of some rule bases.Fuzzy production rule representation method based on rough set theory and forward reasoning mechanism based on Rete algorithm.At the same time,modularity and priority execution control technology are used to resolve rule matching conflicts.(4)Designed and implemented a decision support system for iron ore flotation process.Based on the SSM(Spring+Spring MVC+Mybatis)framework,the overall architecture design of the system has been completed,including the design of flotation databases such as iron ore genes and iron ore types and flotation reagents,rule library design and functional module design,and the development of Java-based Linguistic decision support system.The main functional modules of the system include basic information management module,mineral gene flotation data management module,flotation process decision-making module,flotation reagent consumption prediction module and system management module.In summary,this research has strong practicability and reliability for iron ore flotation prediction and process decision-making.At the same time,it also provides practical guidance for the improvement of iron ore flotation efficiency and control optimization.