THE USE OF KALMAN FILTER IN AUTOMATIC CONTROL OF INDICATORS OF IRON ORES MAGNETIC CONCENTRATION

The paper presents the current state of the problems of automation and control data of wet magnetic concentration for iron ore mining and processing plants. Data on the measurement errors of modern analyzers and material composition of the pulp are considered. The possibility of using Kalman filter in order to obtain the most accurate information about the content of valuable component in concentrate and tailings of magnetic separator products is shown. The mathematical description of the dynamics of concentration indicators is given in state coordinates in the form of differential equations system. The author has selected maximum allowable and nominal values of the water flow in the separator bath and the rotation frequency of its drum as well as the corresponding data for the mass fraction of iron in magnetite concentrate and tailings, based on the reference information. With MATLAB program of computer simulation the nonlinear static characteristics were composed, reflecting the dependence of technological parameters of the magnetic concentration from control actions. The linearization dynamic model of the system is held using expansion in Taylor series in the neighborhood of the points, corresponding to a nominal operation mode. The transfer functions of control valve of water flow rate of induction motor rotating the drum were calculated. Standard deviations of control parameters affecting the separation process were determined. The calculated ratio are obtained for determining the covariance matrices of the system noise, describing dynamics of indicators of magnetic concentration, and the noise of their measurement by devices that control the content of magnetite iron in concentrate and tails. The presented algorithm of estimating the coordinates of the Kalman filter state consists of two stages: prediction of the system state and adjustment of the state vector. Simulation results of the MATLAB programming environment are presented in the form of time diagrams reflecting the dynamics of technological indicators of concentration, the evaluation of the Kalman filter and measurement error. The system has been considered in case of random changes of control actions. At the end of the article, the results are summed up, where it is reported that the Kalman filter should be used in tasks of automation and control of the iron ore concentration process.