Temperature prediction of molten iron transporting process based on extended Kalman filter

Abstract

Molten iron transportation is implemented by torpedo cars controlled by computer-based algorithms, and the process is accompanied by high energy loss, thus it forms a molten iron transportation cyber physical energy system (MIT-CPES). This paper interests in temperature prediction of molten iron transporting process (MITP) derived from MIT-CPES, where the temperature measuring devices locates at the physical layer, historical data of temperature of molten iron, information of torpedo car and weather information are abstracted from information layer. To solve the problem, a nonlinear system model is constructed based on heat transfer mechanism, to describe torpedo car transportation from reladling station (RS) to blast furnace (BF) with residual molten iron. The proposed model is more consistent with the actual production. Moreover, a filter model is also formulated based on extended Kalman filter, in order to predict temperature during the molten iron transporting process. The process data used in the filter model are produced by a multiple linear regression model based actual production data. Numerical results show that the proposed method is superior to traditional mechanism model and beneficial to energy-saving measures in molten iron transporting process.