Subspace Modeling Method for Burn-Through Point

Abstract

In the study, for the iron sintering process which is strongly time-delayed and confounding, a data-driven modeling method called subspace modeling method is proposed to predict the burn-through point (BTP) in this paper. First, by analyzing the mechanism of the sintering process and processing the industrial data, the relationship between the exhaust gas temperature and bellows is confirmed. Then, based on the position of the BTP, a subspace modeling method is used to establish a temperature identification model. This model outputs, the temperature of the BTP and its inputs include the exhaust gas temperature of the front three bellows and the operating parameters. Finally, we compare the obtained result with those of other algorithms. The simulation experiment shows the effectiveness and veracity of the subspace model for BTP, which will provide a precise model for controller design.