Research On The Parameters' Soft Sensing And Synthetic Optimizing Control Of Grinding Process

Being a necessary working procedure in which raw ore is transformed into refining ore, mineral processing industry guarantees the efficient and fluent operation of following smelting procedure. The efficiency of mineral processing mainly depends on the grinding capability of grinding process and recovery rate of floatation process. Between them, grinding process is more important for not only its performance and product's granularity directly influencing the recovery rate of floatation process and the grade of fine ore, but also it occupying the biggest part of resources and cost investment of enterprise. So the operative efficiency of grinding process directly determines enterprise's competitive power and economic profit.Applying automatic control in grinding process is an effective way to enhance the grinding efficiency and product's quality. Regretfully, in the past, owing to the limitations of measuring techniques and control methods, the controlling objectives of grinding operation generally are restrained to external variables of ball mill which are easy to measure, and control methods are usually confined to simple single-circle feedback controls. However, confronting the new situation that the need of grinding product is increasing, the requirement of quality is enhanced, but mining resource is decreasing, traditional measurements and control strategies can hardly satisfy practical demands. Aiming at saving energy, reducing cost and enhancing product's quality, applying advanced control theories and methods to realize the optimizing control of grinding process is a new challenge of research and practical application of control theories.On the basis of deep research on the principle of grinding process, taking realizing synthetic optimizing control to grinding process as objective, the paper presents a systemic scheme for solving the existing problems in grinding control.(1) Grey soft sensor is proposed to conduct real-time measure to grinding concentration, a key parameter of grinding process, for there is no effective way to on-line check the grinding concentration in the past. Uniform incidence degree algorithm is given based on grey analysis theory to select secondary variables of soft sensor, meanwhile, a feasible combination validation method is presented to regularize secondary variables set. Experiment results show that uniform incidence degree algorithm can accurately reflect the relationships...