Study On Optimization Of Ball Mill Operation Of Iron Mine In Inner Mongolia Based On Neural Network

Ball milling operation makes use of grinding medium to impact ore with high frequency to reduce its particle size,providing suitable size and high degree of dissociation for subsequent separation.The mechanism of ball milling determines its characteristics of high energy consumption and medium consumption.The power consumption of grinding operation in iron ore dressing plant of the power consumption of the whole dressing plant.General technical transformation,such as optimization of medium gradation,cannot further achieve energy saving and consumption reduction.Therefore,according to the grinding mechanism and the field technological conditions,this study established the model of energy consumption per ton of ore and operating parameters of grinding operation,and found the optimal operating conditions,so as to achieve energy saving and consumption reduction of enterprises.It is difficult to establish the mechanism model of ball mill because of the strong coupling of operation factors and the high response time delay.Therefore,the energy consumption of tons of ore is taken as the evaluation index,and the data-driven modeling idea based on the input of medium filling rate,mill turning rate,pulp concentration and pellet ratio in ball mill operation is taken as the input.By comparing BP neural network with RBF neural network,the radial basis network is optimized as the framework of modeling according to the site process requirements of iron ore dressing plant.In baiyun obo ore composition complex,to mine dynamic time-varying volatility and the ball mill operation operation condition and induce the forecast model of the problem of lower recognition,RBF neural network structure was optimized by using particle swarm algorithm parameters,avoid choose by artificial experience structure parameters are tentative and blindness,eventually make prediction model fitting error is reduced from 54.6% to 32.9%,plus or minus within 50 KWH of sample points from 7 to 13,the accurate and fast to adjust the scene of the ball mill operation.Finally,through the experimental verification of 105 groups of field data,the ton power consumption of grinding mineral products in ball mill operation was reduced by about 2%,and the occurrence of "overload" and "underload" conditions of ball mill operation was eliminated.This study provides a new approach and intellectual support for the realization of energy conservation,cost reduction and efficiency improvement and the improvement of resource development and utilization level in iron mine.The research idea,prediction model and optimization method of this subject can not only be applied in ball mill operation,but also have wide application in energy saving and consumption reduction in metallurgy,coal chemical industry and other fields,which has certain theoretical value and practical engineering significance.