The Study And Application Of Flotation Production Process Modeling And Control Methods

Paper first discusses the data mining and preprocessing part of flotation processcontrol, including the data mining of offline system and the data preprocessingmethod of online control system. Using Rough Set theory, Fuzzy C-Means algorithm,digital filtering and control diagram method etc. techniques, to achieve theestablishment of the flotation industrial production’s process control model has laid agood, reliable data basis. Considering various variables of input and output offlotation production process, using the regression analysis and principal componentanalysis was carried out on the input data set of flotation to deal with reduced orderprocessing, in order to simplify the input order of RBF Neural Network input model.Use clustering algorithm combined with RBF Neural Network, to achieve the estimateand forecast of flotation production process’s concentrate grade and flotationrecovery’s economic and technical indicators. Then, utilize Rough Set control in theapplication of reverse flotation production research, doing rough rules transformbetween the input pulp grade, size, concentration and flow in the predictive controlmodel, the economic and technical indicators of flotation, and flotation productioncorresponding dosing flow of output pulp, establish an optimized rule table, there is asimple decision-making rule table by dealing with condition attribute and decisionattribute rules. Combined with BP Neural Network, there is the optimized training ofrule base, it has established a medicament flow set point, and then transmits throughthe lower machine to medicament control system, after PI control, it controls thedosage of flotation production exactly. At last, it deals with data by various techniques,and it has achieved the integrated optimization of flotation production finally, it has apositive role on the stability of flotation production and the increase of yield.