Research On Multi-objective Optimization Of Ore Blending Based On Improved Genetic Algorithm

Optimization of ore blending is proportioning the mineral raw material for meeting the requirements of the ore products according to the quality of original ore and mixed ore.Ore blending is an important part in the daily work of ore dressing plant.Qualified ore products can make the dressing work more smoothly.It can improve the service life of the mine and the comprehensive competitiveness of mining enterprises.The research background is ore blending production from one concentrator.According to the actual situation,we can determine the main influence of target.This research has studied the ore blending principle based on the expert experience and has identified the constraints of the optimized ore blending.Selecting the BP neural network algorithm and building mathematical model by data driven.Solution ore blending model by NSGA-II,this research finally obtain reasonable ore blending scheme.The main content of this research are the following:First,the model of ore blending optimization is set up.The existing model usually considers the quality of the ore blending products and ignores the influence of ore distribution to the beneficiation products.In this research,the building model is based on comprehensive concentrate recovery to fit the relationship between the nature of ore blending products and comprehensive recovery as an optimization target by BP neural network algorithm to guarantee the quality of dressing products in the source.In addition,with the passage of time,the available resources in the mine will reduce gradually,so the influence of utilization of mine resources for enterprise should be considered in the blending work.Secondly,starting from the multi-objective optimization problem,this research tells some basic concepts of multi-objective optimization and common methods of multi-objective optimization.This research also introduces an improved multi-objective genetic algorithm(NSGA-II)which has got ideal engineering application on solving some classical multi-objective optimization problems.This study selects this algorithm for solving the optimization model.After getting non-inferior solution set of blending scheme,this research uses TOPSIS to make decisions,and to generate the optimal ore blending scheme.Thirdly,simulation researching on this algorithm by MATLAB,building the software platform by VB and building the database by SQL Server.After all,the optimal ore blending software system is established.The software system is used in the actual production and has brought certain economic benefits for the enterprise.Finally,this research summarizes the full text and puts forward the prospects for the next step of research.