Using of neural control systems in concentration process automation

Abstract

Ball drum mill are widely used on concentrating complexes of iron ore on mining and processing plant. Largely thanks to the beneficial properties of neural networks that are common to different types of nonlinear dynamic objects and solved the problem of identification, analysis, synthesis and hardware implementation of complex process control systems in non-stationary terms and with incomplete and fuzzy information. The implementation of the majority part of considered existing industrial automated control systems are based on proportional, integrating and differential regulators or their combinations (P, PI, PID controllers).However, systems with PID controllers can’t always provide the necessary quality of controling, especially in terms of difficult technological processes with nonlinearity, nonstationarity, delay in time, random perturbations presence of fuzzy and incomplete information. For such technological processes can be attributed most part of mineral processing stages. Milling of iron ore is the preparatory process for the magnetic concentration of materials. The basic technological process of magnetic-concentration complexes is magnetic separation. For the main ways of preparing the ore on magnetic-concentration complexes can be attributed milling and classification of iron ore to controling particle size of ore, that incoming to the magnetic concentration. Considering the large number of components in the composition of the ore and it’s variable characteristics, it is necessary to control of milling processes to achieve optimal size of milling ore and the optimum fractional composition of raw materials that incoming to the magnetic separation.