Study On Grinding Classification Control And Optimization Based On Fuzzy Neural Network

With the development of modern control technology,it is an important way to improve the quality of grinding and classification products by relying on the advanced technology of automatic control to improve and optimize the beneficiation process.However,the grinding classification control process is very complex,and the precision of conventional control algorithm is not high.Aiming at the high-efficiency development and utilization of Shizhuyuan Polymetallic Mine in Hunan Province,this paper studies the optimization of the grinding classification control system of Shizhuyuan polymetallic concentrator.By describing the grinding classification equipment and process control mechanism in the grinding classification control system,as well as the influencing factors and control difficulties of grinding classification control,this paper puts forward the problem of low control accuracy and optimization direction in the grinding classification control system.In this paper,through the research on the current advanced intelligent control algorithm,combined with the actual index requirements of the concentrator,the parameters of each equipment in the grinding and classification process are studied and adjusted.By studying the principles and advantages of fuzzy control and neural network,the two are organically integrated.A grinding classification control system based on fuzzy neural network control algorithm is designed by using MATLAB software.The fitting degree of different algorithms to data is observed by simulating the algorithm.Under the same data,compared with the single neural network algorithm,the fuzzy neural network algorithm has more accurate control accuracy,and can well predict the change trend of pulp overflow fineness.Its actual detection value is closer to the estimated value of the control system,and the error between the two is smaller.Therefore,the fuzzy neural network algorithm can improve the stability and accuracy of the control system,so as to improve the efficiency of the whole grinding and classification system,and further achieve the purpose of energy saving and consumption reduction.Through the man-machine interface of the on-site monitoring system of the concentrator,the real-time operation of the whole grinding classification control system can be understood,and the data report can also be obtained from the monitoring interface.In order to verify the control effect of the fuzzy neural network algorithm in the grinding classification process,the fineness data are analyzed and sorted out according to the particle size distribution of minerals,and the two-stage grinding classification process is simulated by JKSim Met software,The fineness data were mass balanced,and the overflow concentration obtained from the final balance was maintained at 37.54.Compared with the actual value of 35.00,the obtained overflow fineness data is reliable and accurate,and then the data parameters detected and adjusted by the control system can be obtained,which can well and stably control the whole grinding and classification process.According to the working experience of Shizhuyuan field operators and the knowledge of beneficiation experts,a set of control rules are formulated to optimize the algorithm,and the designed fuzzy neural network controller is applied to the grinding and classification system of Shizhuyuan concentrator.By comparing with the previous single neural network control algorithm,the control accuracy of various process parameters in grinding and classification is more accurate.When the given processing capacity is stable,the final overflow concentration remains stable,reducing the working time of the equipment.Thus,the quality of mineral products is improved,and the enterprise can achieve the goal of energy saving and consumption reduction.In the future,when the product quality meets the requirements,the processing capacity of the ball mill can also be appropriately increased to improve the economic benefits of the enterprise.