Velocity Measurement of Blast Furnace Molten Iron Based on Local Multi-Feature Correction Using Multi-Stage Filtered High-Speed Camera

Lei He,Zhaohui Jiang,Zhipeng Chen,Weihua Gui,Yongfang Xie

DOI: https://doi.org/10.1109/jsen.2020.2997862

IF: 4.3

2020-01-01

IEEE Sensors Journal

Abstract:Accurately measuring the velocity of molten iron at the taphole of a blast furnace is of great significance for safe tapping and reflecting the operating state of the blast furnace. The presence of dust-induced brightness changes and weakly textured areas on the surface of molten iron seriously affects the accuracy of the surface velocity field reconstructed by traditional methods. This paper presents a local multi-feature correction using multi-stage filtered high-speed camera to achieve high-precision measurement of the surface velocity field of molten iron. First, two sets of local molten iron features were extracted according to a theoretical analysis of the influence of dust-induced brightness changes and the molten iron texture on the accuracy of combined local-global approach with total variation regularization (CLG-TV) and digital image correlation (DIC). Then, a mixed semifuzzy approximation network based on the characteristics of local molten iron features is proposed to estimate measurement errors under the influence of brightness changes and weak textures. Finally, an ingenious decision maker is designed to determine a more accurate displacement field of the molten iron surface. Additionally, a designed multi-stage filtered high-speed camera is installed beside the taphole to provide good optical images of molten iron surface. Industrial experiments and applications illustrate that the proposed method can stably and accurately measure the molten iron velocity at the taphole and provide reliable guidance for tapping operations.