Flow of molten slag and iron at 1500 °C to 1600 °C through packed coke beds

• Published in: Metallurgical and materials transactions. B, Process metallurgy and materials processing science Vol. 36; no. 6; pp. 765 - 776
• Main Authors: HUSSLAGE, W. M, REUTER, M. A, HEEREMA, R. H, BAKKER, T, STEEGHS, A. G. S
• Format: Journal Article
• Language: English
• Published: Heidelberg Springer 01.12.2005; Springer Nature B.V
• Subjects: Applied sciences; Coke; Exact sciences and technology; Fluid dynamics; Iron; Metallurgy; Metals. Metallurgy; Production of metals; Slag; Temperature; Slag; Coke
• ISSN: 1073-5615; 1543-1916
• DOI: 10.1007/s11663-005-0080-6

The blast furnace dripping zone is of great importance to the mass transfer of elements such as sulfur, carbon, and silicon, to and from the liquid metal phase. To understand mass transfer in the dripping zone, not only mass-transfer reactions and kinetics should be known, but the flow phenomena and process dynamics should be understood as well. The flow of hot metal and slag in the dripping zone was studied in experiments, in which liquid slag and metal trickled through a packed coke bed at 1500 °C to 1600 °C. The results indicate that slag and iron flow concurrently in a funicular type of flow. The iron flows through the core of the voids in the bed and is enveloped by slag, which flows filmwise in between the coke and the iron. This mode of flow allows for a large contact area between slag and iron, through which mass can be transferred. While flowing, the liquid can only pass and access a void, if and when the fluid capillary pressure at the void neck can be overcome. As a result, liquid droplets collect into rivulets. These rivulets flow down, along the accessible voids, using only a part of the available volume. The residence times of the fluids in the bed depend partly on the length of the pathway and are a function of the bed structure, the void neck distribution, and the stochastics of the flow. During flow, slag may react with coke, thus changing the distribution of the slag composition, and its sulfur capacity. In addition, the residence time distribution of the slag and the liquid holdup change as a result of these reactions. Holdup and residence time distribution of the liquids as measured in the experimental setup could not be modeled quantitatively, most likely due to the doubly distributed nature (in space and in time) of the model parameters, induced by reactions between slag, coke, and liquid metal. [PUBLICATION ABSTRACT]