Research On Dynamic Measurement Of Polymetallic Ore Cut-off Grade And Ore Blending Optimization Technology

With the development of intelligent mines,artificial intelligence and computer technology are increasingly being successfully applied in mines.The dynamic measurement of cut-off grade and intelligent blending of polymetallic ore is a complex multi-objective optimization problem,involving geology,mining,mineral processing,mathematics,artificial intelligence and computers,etc.The research scope is wide and complex,and the dynamic optimization of cut-off grade and intelligent blending is an important way to improve the competitiveness of mines,and its research prospect is broad and meaningful.This thesis takes complex low-grade polymetallic ore as the object,on the basis of analyzing the connotation of mine techno-economic index system and the complex dynamic correlation among its indicators,introducing artificial intelligence and computer technology,studying the multi-objective optimization method of mine operation parameters based on hybrid particle swarm and genetic algorithm,and realizing the real-time,dynamic and overall optimization of cut-off grade and ore blending under complex supply conditions.The main conclusions are as follows.1.The complex linkage relationship between the cut-off grade and reserves,depletion rate,loss rate,mineral processing recovery,mining cost and mineral product price is researched.The particle swarm algorithm is introduced to solve the dynamic optimization of polymetallic ore cut-off grade,and the particle swarm algorithm is improved accordingly to realize the multi-objective optimization of polymetallic ore cut-off grade,which provides a new solution idea for the dynamic measurement of polymetallic ore cut-off grade.2.A new method of using genetic particle swarm algorithm to solve the dynamic optimization of cut-off grade is proposed.The linkage relationship among four parameters of reserves,mining grade,cost and price of polymetallic mines is studied,and a model for solving the interlinkage of the four parameters is established.The method can quickly calculate the profit and loss grade limit,cost to price ratio and reserve availability of polymetallic mines,which provides a scientific means for the implementation of reserve dynamic management,cut-off grade dynamic estimation and optimal ore blending in polymetallic mines.3.A multi-objective ore blending optimization method based on hybrid particle swarm optimization algorithm is proposed,and a multi-objective model combining short-term production planning and dynamic ore blending optimization in mines is established,and the basic particle swarm algorithm is integrated with the crossover and variation operations in the standard genetic algorithm.The optimization speed and precision of hybrid particle swarm optimization algorithm are improved.4.A hybrid particle swarm algorithm for solving the ore blending optimization model,the ideal point method,is studied.The ore output grade constraint and ore mined quantity constraint in the constraints are transformed into the objective functions of monthly ore output grade and monthly ore mined quantity,respectively,which are taken as sub-objective functions.Due to the many variables involved in the ore blending optimization model,which leads to a large search space range of constraint functions and increases the difficulty of solving the optimization model,the improved ideal point method proposed in this thesis accelerates the particle search speed by adjusting the particle advance direction,and its convergence speed is fast and its robustness are high.5.The particle search strategy with kernel particles and double attractors is proposed.The iterative process of the ordinary particle swarm algorithm will have multiple non-inferior solutions,which leads to the increasing size of the external archive and the decreasing threshold,making the particle swarm more and more concentrated in the local search region,thus,increasing the difficulty of the particle swarm search for the initial feasible solution.However,for optimization problems with complex constraints,the determination of the initial feasible solution is very difficult,and it is again difficult for particles searching outside the feasible solution region to enter the feasible solution region.By improving the search coefficients of the particle swarm algorithm and adopting the particle search method with kernel particles and double attractors,the global search capability of the algorithm can be enhanced to solve the optimization problem of ore blending under complex constraints.The cut-off grade dynamic measurement method proposed in this thesis is an improvement and enhancement of the traditional grade calculation method,and the corresponding computer software system is successfully developed to realize the organic combination of cut-off grade dynamic testing and ore blending,which greatly improves the speed and accuracy of cut-off grade measurement and the quality and efficiency of ore blending management in polymetallic mines.The cut-off grade dynamic optimization method,ore blending technology and software system proposed in this thesis have been successfully applied in Chengmenshan copper mine of Jiangxi Copper Company and Chongqing branch of Aluminum Corporation of China,which have gained good economic and social benefits.