Soft Sensor Modeling of Blast Furnace Wall Temperature Based on Temporal–Spatial Dimensional Finite-Element Extrapolation

Abstract

The blast furnace (BF) wall temperature field plays a key role in determining whether the operation of the BF is efficient. However, it cannot be measured directly because the BF is an enclosed complex reactor with an extremely harsh environment. To cope with this problem, a soft sensor modeling strategy of the BF wall temperature field based on a novel temporal–spatial dimensional finite-element extrapolation algorithm (TS-FEEA) is designed. A 3-D transient heat conduction model of the BF wall is first established, which considers the layered physical structure of the BF. The large volume of the BF wall causes great calculation burden when solving the 3-D transient heat conduction model of the BF wall, to solve this problem, a TS-FEEA is proposed that realizes a high-speed and accurate model numerical solution. To further improve the 3-D model precision, the measured thermocouple temperatures in the cooling stave are served as the judgment criteria, and a correction strategy for inner wall temperature is finally designed to correct the assumed inner wall temperature through multiple iterations. Simulation results demonstrate the validity of the soft sensor modeling strategy and its applicability in industrial plants. The research achievements provide valuable reference information for the BF operators to judge the furnace conditions and adjust the operation parameters.