Predictive control of the iron ore beneficiation process based on the hammerstein hybrid model

Abstract

Non-linear, dynamic, non-stationary properties characterize objects of the iron ore beneficiation line. Therefore, for their approximation, it is advisable to use models of the Hammerstein class. As a result of comparing the three models of Hammerstein: simple, parallel and recursive-parallel, it was shown that the best result for identifying the considered processes of magnetic beneficiation of iron ore by the minimum error criterion was obtained using the Hammerstein recursive-parallel model. Hence, it is recommended for the identification of beneficiation production objects.