The Automatic Control System Of Flotation Dosing Based On Fuzzy Expert Reasoning And Data Mining Technology

Flotation automatic dosing has been used in mineral processing plants for many years,which can remotely and automatically control and adjust the accurate amount of chemical added,overcoming the various adverse effects caused by the failure of flotation operators to timely and effectively detect changes in the flotation process and adjust them.However,at present,the automatic control system of flotation dosing has the problem that the algorithm is simple and it is difficult to accurately adapt to the changes of feed ore or track product indicators,and this project studies a flotation dosing automatic control system based on fuzzy expert reasoning and data mining technology,which solves the problem that the dosing system cannot be adjusted in real time according to the flotation feed change,flotation state change and index fluctuation in the flotation process.According to the anti-flotation process and operation control factors of Qi Dashan beneficiation plant,combined with the characteristics of the flotation process of the beneficiation plant,a set of flotation dosing automatic control system based on fuzzy expert reasoning and data mining technology is developed.This system establishes flotation control factors,state data and flotation index relationship rules such as flotation feed amount,slurry concentration,slurry p H,flotation tank aeration capacity,flotation tank stirring speed,foam layer thickness,foam color,foam stability,etc.,through small-scale flotation test and on-site production collection information,establishes a database,summarizes the algorithm rules such as concentrate grade prediction and dosing amount control of the dosing automatic control system,and can accumulate and train through long-term data accumulation.Achieve increasingly accurate prediction and control.The information collection in the system is mainly divided into two parts,one is to extract foam texture feature information by using the symbiotic grayscale matrix algorithm,and the other part is the monitoring data.All collected information is stored in a database.The data were preprocessed by data mining technology for rough set attribute reduction algorithm and PCA principal component analysis method.The GRNN neural network prediction model is established to predict the iron grade of concentrate,and the absolute error can be controlled within 2%.According to the error between the predicted iron grade of concentrate and the target value of the required grade of the production requirement,a fuzzy expert system was used to guide the amount of agent added.The system combines on-site monitoring data,designs database types,and uses SQLite software to establish databases to facilitate data call and access;The data were preprocessed by Matlab software using rough set attribute reduction algorithm and PCA principal component analysis method,and the K-fold cross-validation method was used to combine with GRNN neural network to construct a GRNN neural network prediction model,which was used to predict the iron grade of concentrate,establish a fuzzy expert rule base,fuzzy processing the input and output data,and use the fuzzy inference machine to complete the acquisition of the amount of chemical added.Finally,in the C++ development environment,the interface of "current production process control index","current production chemical index" interface and "concentrate grade display" interface were established to complete the development of flotation dosing automatic control system.The system fully considers the influence of various control factors on the flotation index in the flotation process,reasonably predicts the concentrate grade and flotation agent dosage according to the flotation collection data,and establishes a set of flotation dosing automatic control system suitable for industrial sites based on this research,which is applied to the flotation process of Angang Qi Dashan second selection workshop,which is close to the normal judgment level of manual labor,and basically realizes the prediction of concentrate grade and the guidance of chemical flow.