Study On Optimization And Semi Physical Simulation System Of Nickel Flotation Process

Flotation,as a method of changing the physical and chemical properties of mineral surface through flotation reagents,is of great significance to industrial beneficiation.For a long time,it is difficult to control the grade of flotation concentrate because of long flotation process,complex flotation process and many flotation control parameters.Based on the industrial application,according to the difficulties in flotation process modeling and flotation control parameter optimization,this paper puts forward specific solutions and improvement measures.The main work is as follows:(1)The hardware in the loop simulation system of nickel ore flotation is introduced: through the analysis of flotation principle,the flotation process of the system is further understood,so as to analyze the relationship between the working condition parameters on the flotation column and the target grade,and point out the influence of the working condition parameters on the grade change.On the basis of these two steps,this paper systematically expounds the realization mode and construction process of the hardware in loop simulation system of nickel ore.(2)To deal with the fault errors and random errors in the process of data acquisition: to eliminate the fault errors in the process of data acquisition,and according to the characteristics of random errors,a new filtering idea combining sliding window filtering,pauta criterion and arithmetic average filtering is proposed to denoise;and to deal with the process parameters and large fluctuation range,the module In the case of imprecise type,K-means clustering is used to separate different working conditions,and for the case that the clustering results are not repeatable,this paper introduces the concept of density,improves K-means,and finally makes the clustering effect stable and repeatable.(3)In view of the long process of flotation system and the mutual influence of each link of flotation,this paper proposes that the parameters of flotation process are time series,and chooses NARX to build the model of this system;in view of the classification of NARX structure and workingcondition,this paper proposes the method of data splicing to splice the data of different time periods but the same type,avoiding the same time In this paper,the momentum conjugate gradient method is used to update the weights and thresholds of the neural network,and good results are achieved.(4)In view of the fact that there are many control parameters in the flotation system,and the control parameters affect each other to determine the flotation column grade,so it is difficult to solve the problem of the best control parameters to improve the concentrate grade combination by using the classical optimization algorithm.In this paper,the swarm intelligence algorithm is used to find the best control combination.This paper introduces the particle swarm optimization.Furthermore,in this paper,aiming at the disadvantage that the constant flying speed of particle swarm optimization algorithm may lead to the particle flying over the best value in the optimization process,the inertia weight is updated by the way of waveform decreasing,and the algorithm is better than the particle swarm optimization algorithm in the optimization effect.(5)According to the actual engineering needs,taking MATLAB as the platform,on the premise of analyzing the software requirements,a software simulation platform is built to set up the modeling,control parameter optimization and real-time simulation,which can build the model in real time,optimize the control parameters according to the current working conditions,and predict the change trend of grade according to the optimized control parameters and working conditions,and graph the trend to improve Improve the friendliness of human-computer interaction.