外文文獻(xiàn).pdf

Estimation of carbon balance in reaction zones of a submerged-arc furnace during ferrosilicon smelting Bolesaw Machulec1,a, Grzegorz Kope1,b 1 Faculty of Material Science and Metallurgy, Silesian University of Technology, Krasiskiego 8, 40-019 Katowice, Poland a boleslaw.machulecpolsl.pl, b grzegorz.kopecpolsl.pl Keywords: ferrosilicon, carbon balance, submerged-arc furnace. Abstract. Based on the electrical parameters of a 20 MVA ferrosilicon furnace, a methodology of identification of characteristic ferrosilicon smelting states described as carbon excess and carbon deficiency in the reaction zones (over-coked, under-coked) has been presented. Relations between the electrical parameters and assessments made by furnace operators regarding characteristic furnace states related to amounts of carbon in the reaction zones have been demonstrated. The results show that the reactive power measurements as well as the k, c3 (Andreaes, Westlys) parameters, provide the same information on the furnace state and have a close relation with resistivity of the current-conducting zones. Similar information on the level of carbon balance in the reaction zones is obtained from the harmonic analysis of phase voltages and currents or measurements of higher harmonic components using high-pass filters. Introduction Ferrosilicon is smelted in electric submerged-arc furnaces with Sderberg self-baking electrodes 1-3. Ferrosilicon smelting is a continuous process. Raw materials are intermittently loaded into the furnace from the top as a charge mix that contains: quartzite, carbon reducers (hard coal, pea coke and wood chips) and ferriferous materials (roll scale or steel chips). Cyclically (approximately every 2 hours), liquid metal is transferred to the ladle during tapping through one of the tapholes located near the hearth in the side wall of the furnace bath. Heat that is necessary for the conduct of strongly endothermic reaction of carbon reduction of silica is released due to direct flow of electrical current through the charge material (resistive heating) and through the electric arc that burns in the gas chambers located near the electrode tips. When simplifying and assuming a 100% iron yield, the 75FeSi ferrosilicon smelting process can be described by means of the following overall reaction 1-3: ?+ ? + 0.167? = ? + 0.167? + ?1 ? + ?(g) (1) where: ? 1 + ? (2) ? ?0.80 0.94?. Gaseous silicon oxide, ?, is one of intermediate products during the electrothermal process of ferrosilicon smelting that is formed in the near-electrode tip furnace zones of the highest temperature as a result of the following reaction: 2SiO2 + SiC = 3SiO(g) + CO(g). (3) Simultaneously with the reaction (3), other reactions that produce metallic silicon occur: ? + ? = 2? + ?(g) SiO2 + 2SiC = 3Si + 2CO(g) . (4) Solid State Phenomena Vol 226 (2015) pp 11-16Submitted:2014-05-31 (2015) Trans Tech Publications, SwitzerlandRevised:2014-09-30 doi:10.4028//SSP.226.11Accepted:2014-10-01 Online:2015-01-12 All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans Tech Publications, . (ID: , California Institute of Technology, Pasadena, USA-27/05/15,16:24:37) Gaseous silicon oxide, ?, can also be generated in the arc gas chambers during the following reaction: ? + ?= 2?. (5) Gaseous ?, resulting from the reactions (3) and (5), passes upwards and reacts with carbon in the furnace zones of lower temperature: SiO(g) + 2C = SiC + CO(g). (6) In addition to the reaction (6), other reactions may also occur that lead to deposition of gaseous ? oxide, (it undergoes condensation), in the charge material of the upper furnace zones: SiO(g) + CO(g) = SiO2(g) + C 3SiO(g) + CO(g) = 2SiO2 + SiC 2SiO(g) = Si + SiO2 ? + ? = ?+ ? . In solid state, ? passes downwards with partly reacted charge material to the zones of the highest temperature where it is used in the reactions (3) and (4) resulting in ? gas is generated and silicon is produced. Reaction (6), which leads to ? formation, may also occur in the arc zones as not all amounts of carbon delivered with the charge can entirely react in the upper zones of the furnace. The physicochemical conditions for the reaction (6), (7) which leads to ? formation, have a significant impact on the silicon yield, and efficiency of the ferrosilicon smelting process. The reaction (6) of ? condensation becomes slower during deposition of ? on the reducer surface and its course is markedly affected by reducer reactivity 1-2. Part of ?, which is not subjected to condensation, reacts with atmospheric oxygen when leaving the furnace and forms microsilica that is one of ferrosilicon smelting by-products. If not all amounts of ? are used in the reactions (3) and (4), its excess may deposit in the furnace working space, which is related to excess amounts of the reducer in the reaction zones. The excessive amounts of ? may deposit in the lower zones of the furnace as an infusible and non-reactive allotropic form, ?, that results from the following transformation: ? = ? (8) The effects of this are: technological process disturbances, metal tapping problems and, in extreme cases, long-term furnace capacity reduction. Therefore, the practical goal is to perform the technological process with slight reducer deficiency so that all amounts of ? are used and do not deposit in the furnace. Carbon balance in the furnace working space is difficult to determine and cannot be directly measured. It is markedly affected by the amounts of reducer in the charge material; however, chemical compositions of reducers are subjected to uncontrolled, random fluctuations. Despite a significant progress of process automation and introduction of technology support systems, furnace states related to excess or deficiency of carbon in the reaction zones (over-coked, under- coked) are mostly determined by furnace operators based on the electrical measurements and observations of the furnace operation course. These assessments depend on perceptivity and experience of the immediate furnace operators and are highly subjective. Lack of consistent assessment criteria for carbon balance in the reaction zones results in temporary disturbances of the technological process, which negatively affects the capacity as well as technical and economical indicators. Therefore, development of an objective method of the assessment of characteristic furnace states regarding excess and deficiency of carbon in the reaction zones is important for improvement of stability as well as technical and economical indicators of the ferrosilicon smelting process. (7) 12Technologies and Properties of Modern Utility Materials XXII Identification of carbon balance in the ferrosilicon process Carbon-containing components of the furnace working space has a much lower electrical resistivity as compared with quartz 1 and therefore, their contents significantly affect bath resistance and the current density distribution in the ferrosilicon furnace. As a result, amounts of carbon in the working space of the furnace are related to the stoichiometric parameters process as well as electrical, thermal and temperature properties of the reaction zones. Electrical measurements of the furnace are the basis for identification of its characteristic states defined as over-coked and under- coked condition in the reaction zones. For a long time, the reactive power has been a major criterion for decisions on corrections of reducer amounts in the charge material burden. Moreover, important parameters, known from many publications, are k, ? (Andreaes, Westlys) which describe electrical and thermal similarities of the reaction zones furnaces of various parameters 3-7. Another approach is identification of amounts of power generated in the electric arc, considering its non-linearity as an element in the furnace electrical circuit. The basis is a harmonic analysis of phase voltages and currents or measurements of higher harmonic components using high-pass filters 6, 8. Based on the electrical parameters of a 20 MVA furnace with Sderberg electrodes, d = 1.2 m, a methodology of identification of characteristic states during ferrosilicon smelting related to the amounts of carbon in the reaction zones were presented. The electrical parameters were recorded at 1-minute intervals by the furnace computer-controlled measurement system. Simultaneously, assessments made by the furnace operators were analysed in relation to the states of the process accompanied by changes in the reducer content in the charge material burden. Subjects of the analysis were 24-hour periods when the furnace operated regularly without breaks or power reduction. In Fig.1, examples of the furnace electrical parameter changes versus time are presented. In the figure, the times of reducer content changes in the charge material burden in relation to identification of the characteristic furnace states by the operators were marked. The relations, presented in Fig. 1, describe one of the investigated furnace operating periods which properly determines a relation between the electrical parameters and the states over-coked, under- coked identified by its operators. The results, presented in Fig. 1, show that the reactive power measurements as well as the ?,? (Andreaes, Westlys) parameters, provide the same information on the furnace state and have a close relation with electrical resistivity of the current-conducting zones. Similar information is obtained from the relation that presents the coefficient of higher harmonic components in the phase voltages 7 (Fig. 2). It should be noted that the electrical parameters of the furnace bath cannot be only identified with the Cfix content, in the charge material burden. The electrical parameters of the furnace bath are strongly affected by its inner structure particularly by thickness of carbide walls that form gas chambers of the arc and by the amounts of infusible SiC that is deposited in the furnace. The inner structure of the furnace bath or the reaction zones cannot be considered as time-lasting. Depending of a set of technological factors, individual parts of arc gas chamber walls increase or decrease or even disappear. Due to this, temporary changes in reducer content in the charge not always immediately affect the electrical properties of the furnace bath. However, when the technological process is adequately controlled, the relations presented in Figs. 1 and 2 can be used in the automatic systems of identification of characteristic furnace states related to the control of carbon amounts in the reaction zones. This should, to some extent, allow for neglecting subjective assessments of the furnace state made by its operators regarding of carbon balance. The relations presented in Figs. 1 and 2 refer to the furnace operating period which, due to the technical and economical indicators of raw material and energy consumption, was not “perfect” (E = 8810 kWh/Mg pure metal). Ranges of the parameters corresponding to the carbon balance state of the reactions zones can be determined based on the data from the periods when the furnace operation is “perfect”, i.e. it shows maximum capacity and the best technical and economical indicators. In Fig. 3, courses of parameters of carbon balance state in the investigated 20 MVA furnace are presented. Solid State Phenomena Vol. 22613 + reducer 300 - 305 - reducer 305 - 300 0.80 0.90 1.00 1.10 1.20 6:0010:0014:0018:0022:002:006:00 c3,AW-2/3 +reducer 300/305 - reducer 305/300 0.01 0.02 0.03 0.04 0.05 0.06 6:0010:0014:0018:0022:002:006:00 Fig. 1. 24-hour changes in the reactive power (MVar) and the electrical parameters of furnace reaction zones (k Andrea, c3 Westlys parameters) 8, (FeSi75, 20 MVA furnace, Burden kg: quartzite 500, coal 300305, scale 82, chips 100, energy consumption indicator E = 8810 kWh/Mg) Fig. 2. The coefficient of odd harmonic components of phase voltages versus time and changes in carbon amounts in the burden 8 (FeSi75, 20 MVA furnace, Burden kg: quartzite 500, coal 300-305, scale 82, wood chips 100, E = 8810 kWh/Mg) + reducer 300 - 305 - reducer 305 - 300 7 8 9 10 11 12 13 6:0010:0014:0018:0022:002:006:00 MVar + reducer 300 - 305 - reducer 305 - 300 4.E-03 5.E-03 6.E-03 7.E-03 8.E-03 6:0010:0014:0018:0022:002:006:00 k, m 14Technologies and Properties of Modern Utility Materials XXII Fig. 3. Parameters of carbon balance state in the reaction zones, (FeSi75, 20 MVA furnace, Burden kg: quartzite 500, coal 293-309, scale 72, wood chips 120, E = 8105 kWh/Mg pure metal) Summary The amounts of carbon in the furnace working space are related to stoichiometric parameters of the process as well as electrical, thermal and temperature properties of the reaction zones. Lack of consistent assessment criteria for carbon balance in the reaction zones is a reason of not always appropriate decisions taken by ferrosilicon furnace operators. This results in temporary ferrosilicon smelting technological process disturbances, which negatively affects the capacity as well as technical and economical indicators. Electrical measurements of the furnace are the basis for identification of its characteristic states defined as excess and deficiency of carbon in the reaction zones (over-coked, under-coked). The study results show that electrical measurements: reactive power and the k, c3 (Andreaes, Westlys) parameters that describe thermal and electrical similarities of the reaction zones provide the same information on the furnace states related to excess and deficiency of carbon. Comparable information is obtained from the harmonic analysis of phase voltages and currents as well as from identification of the electric arc effects on heat release in the furnace. The electrical measurements enables automated identification of the furnace states related to amounts of carbon in the reaction zones. This, to some extent, allows for neglecting subjective assessments made by the furnace operators. 4.3 4.8 5.3 5.8 6.3 22:002:006:0010:0014:0018:0022:00 k, m m Time 0.90 0.94 0.98 1.02 1.06 22:002:006:0010:0014:0018:0022:00 C3, AW(-2/3) Time 8 9 10 11 12 22:002:006:0010:0014:0018:0022:00 MVar Time Solid State Phenomena Vol. 22615 References 1 A. Schei, J.Kr. Tuset, H. Tveit, Production of High Silicon Alloy, Tapir Forlag, Trondheim 1998. 2 B.F. Lund, Rigorous simulation models for im