锌冶金 2005 - Optimization of rotary kiln treating EAF dust

Journal of University of Science and Technology Beijing Volume 12, Number 4, August 2005, Page 308 Corresponding author: P e n g f u T a n , E-mail: ta n p e n g f u @yah o o . c o m Metallurgy Modeling and optimization of rotary kiln treating EAF dust Pengfu Tan and Pierre Vix M o u n t I s a M i n e s L i m i t e d , M O U N T I S A , Qu e e n sl a n d 4 8 2 5 , Aus t r a l i a ( R e c e i v e d 2 0 0 4 - 0 9 - 3 0 ) Abstract: Electric arc furnace (EAF) dust from steel industries is listed by the United Sates EPA as a hazardous waste under the regulations of the Resource Conservation and Recovery Act due to the presence of lead, cadmium and chlorine. The disposal of the approximately 650000 t of EAF dust per year in the U.S. and Canada is an expensive and unresolved problem for the majority of steel companies. The Waelz process has been considered as the best process for treating the EAF dust. A process model, combined thermodynamic modeling with heat transfer calculations, has been developed to simulate the chemical reactions, mass and heat trans- fer and heat balance in the kiln. The injection of air into the slag and the temperature profile along the kiln have been modeled. The effect of (CaO+MgO)/SiO2 on the solidus temperature of slag has also been predicted and discussed. Some optimized results have been presented. Key words: thermodynamic model; EAF dust; rotary kiln; solidus temperature 1 Introduction In recent years, electric arc furnace (EAF) technol- ogy has emerged as a significant segment of the steel industry. The dust, generated during the operation of the electric arc furnace, has been listed as hazardous solid waste by the Environmental Protection Agency. The lead, cadmium and chromium contained in the EAF dust are considered hazardous. The EAF dust compositions from any given plants vary widely primarily due to the fluctuating in its feed (steel scrap) chemistry. The typical chemical analysis of EAF dust (wt%) is: zinc, 18; lead, 2.3; cadmium, 0.05; chlorine, 1.7; and fluorine, 0.5. Waelz kiln [1-6] is the best available techniques (BAT) to treat EAF dust. The EAF dust together with coke breeze or coal as reluctant are simultaneously proportioned and charged into the rotating kiln. One burner and coal combustion heat the kiln and the bed. Zinc, lead and cadmium volatilize as metal vapor and are re-oxidized in the kiln atmosphere. The kiln has to be operated with surplus air in order to produce metal oxides. The mixed oxides are drawn from the kiln with the flue dust and separated in a gas-cleaning system. The slag leaves the kiln at the opposite end and is granulated. 2 Dynamic model of rotary kiln Thermodynamic modeling has been used to simu- late copper smelting [7], lead smelting [8] and nickel smelting [9-10] in a number of studies and proved to be very successful. The author and co-worker have also developed several thermodynamic models to simulate the copper smelting process [7-8], direct lead smelting process [9] and nickel smelting process [10- 12] and dioxin formation in iron ore sintering process [11-12]. The two reaction zones of the kiln, the gas zone and bed zone, are schematically shown in figure 1. A dy- namic model, combined thermodynamic calculations in each cell with the calculation of heat transfer be- tween cells, has been developed to simulate the chemical reactions, and mass and heat transfer in the kiln. This model is based on the previous work done by Koukkari and Penttilä [13-14]. In the model, the cells in the two zones are assumed to be controlled thermodynamically. Thermodynamic data for the equilibrium calculations in each cell are extracted from FactSage thermodynamic database [15]. Figure 1 Thermodynamic calculation cells inside the kiln. The heat transfer includes convection and radiation from gas to the bed, convection and radiation from gas to the inner wall, conduction and radiation from the P.F. Tan et al., Modeling and optimization of rotary kiln treating EAF dust 309 inner wall to the bed, conduction from the inner wall to the outer wall, convection and radiation from the outer wall to surroundings. In figure 2, Qgb is convec- tion and radiation heat transfer from gas to the bed; Qgi is convection and radiation heat transfer from gas to the inner wall, Qib is conduction and radiation heat transfer from the inner wall to the bed; Qio is conduc- tion heat transfer from the inner wall to the outer wall; and Qos is convection and radiation heat transfer from the outer wall to surroundings. Figure 2 Heat balance calculation cells inside the kiln. The heat transfer from gas to the bed is: 4 4 gb gb gb g b b g b( ) ( )Q h A T T GS T Tσ= − + − . The heat transfer from gas to the inner wall is: 4 4gi gi gi g i i g i( ) ( )Q h A T T GS T Tσ= − + − . The heat transfer from the inner wall to the bed is: 4 4 ib ib ib i b i b i b( ) ( )Q h A T T S S T Tσ= − + − . The heat transfer from the inner to the outer wall is: i o io 1 2π( ) 1 ln n j j jj T TQ r rλ + −= ∑ . The heat transfer from the outer wall to surroundings is: 4 4 os os os o s o s o s( ) ( )Q h A T T S S T Tσ= − + − . where r is the radius, A the heat transfer area, h the heat transfer coefficient, T the temperature, GS the total exchange area between gas and surface, and 1 2S S the total exchange area between two surfaces (bed and wall). The counter-current streams are divided into volu- me elements that exchange heat and matter with each other and the surroundings. The streams encounter each other in a "zipper" iteration that converges ac- cording to the outgoing and incoming temperatures that are also known by measurement. The overall heat balance is checked by including the burner in the multi-component thermodynamic model. 3 Results and discussion Figure 3 shows the calculated temperature profiles of the gas, bed, inner wall and outer wall in the kiln. The temperature in the feed end is much higher than that in the slag discharge end. In general, the tem- perature of gas is higher than that of slag. Figure 3 Predicted temperature profiles of the gas, bed, inner wall and outer wall in Waelz kiln. Figure 4 presents the amounts of slag, coal, limestone and water along the length of the kiln. The water is vaporized in the first 6 m of the kiln from the feed end. The limestone decomposes completely at the middle of the kiln. In general, some unreacted coal 310 J. Univ. Sci. Technol. Beijing, Vol.12, No.4, Aug 2005 remains in the discharged slag in order to reduce all zinc and lead oxides. According to figure 4, 30% of all coal charged has not reacted with the EAF dust, and remained in the slag. The amount of the discharged slag is only half of the amount of the initial feed. Figure 4 Amounts of solids along the length of Waelz kiln. 3.1 Effect of coal charge Table 1 shows the effect of coal charge on the op- erations of the Waelz kiln. The zinc recovery in- creases as the coal charge increases. But the unreacted coal in the slag also increases as the coal charge in- creases. Figures 5 and 6 present the effect of coal charge on the temperature profiles of gas and bed along the length of the kiln. When the percentage of carbon in the feed is greater than 15%, the amount of coal has very little influence on the temperature profiles of the gas and bed along the length of the kiln, as shown in figure 5. But the amount of the remaining coal in the slag increases significantly, according to table 1. It means that the extra coal does not react with the EAF dust. Under this operating condition of Waelz kiln, 15% of carbon in the feed is enough to reduce most of the zinc and lead oxides. The temperature of the gas and bed decreases along the length of the kiln, and the recovery of zinc de- creases from 88.8% to 76.5%, as the percentage of carbon in the feed decreases from 15% to 12%. When C is only 11wt%, the kiln cools, and the kiln tem- perature is so low that the oxides can not be reduced, as shown in figure 6 and table 1. Table 1 Effect of coal charge on the operations of Waelz kiln C in feed / wt% Discharged slag temperature / °C Off-gas temperature / °C Unreacted C in slag / wt% Zn in slag / wt% Zn recovery / % 20 863 831 11.9 3.9 91.1 17 858 834 5.9 6.3 86.3 15 860 846 4.4 5.0 88.8 12 857 815 0.2 11.0 76.5 11 730 565 4.0 30.9 0 3.2 Effect of second air injection Figure 7 presents the predicted temperature pro- files of the gas, bed, inner wall and outer wall in the Waelz kiln when air is injected under low pressure at the discharge end of Waelz kiln. Temperatures of the bed and gas in the feed end of Waelz kiln with second air injection are much lower than that of normal Waelz kiln, but temperatures of the bed and gas in the discharge end are much higher than that of normal Waelz kiln, as shown in figures 3 and 7. The reason is Figure 5 Effect of extra coal charge on the tempera- ture profiles of gas and bed in Waelz kiln. Figure 6 Effect of coal charge on the temperature profiles of gas and bed in Waelz kiln. P.F. Tan et al., Modeling and optimization of rotary kiln treating EAF dust 311 that the metallic iron in the slag re-oxidizes with the blast air, and the oxidization reaction releases a sig- nificant amount of heat near the discharge end of the Wealz kiln where the second air is injected. Figure 7 Predicted temperature profiles of the gas, bed, inner wall and outer wall in Waelz kiln where the second air is injected. Figure 8 shows the effect of coal charge on the temperature profiles of the gas and bed along the length of the Waelz kiln where the second air is in- jected. When carbon in the feed decreases from 15% to 10%, the temperature of the gas and bed are still high enough to reduce the zinc and lead oxides. Figure 8 Effect of coal charge on the temperature profiles of gas and bed in Waelz kiln where the second air is in- jected. 3.3 Solidus temperature of the slag and CaO/SiO2 in feed In order to avoid accretion problems in the kiln, the solidus temperature of slag should be high enough. The CaO/SiO2 or (CaO+MgO)/SiO2 ratio is very im- portant for the control of solidus temperature of slag. The soildus temperature of the slag has been cal- culated using FactSage [15] and shown in the follow- ing figures. Figure 9 shows the effect of CaO/SiO2 ratio on the solidus temperature of slag from the feed end to the middle of kiln. Figure 10 presents the ef- fect of CaO/SiO2 ratio on the solidus temperature of slag from the middle of kiln to the discharge end. Figure 9 Solidus temperature of slag from the feed end to the middle of kiln. Figure 10 Solidus temperature of slag from the middle of kiln to the discharge end. 3.4 Solidus temperature of slag and MgO content in feed Figure 11 shows the effect of the replacement of CaO by MgO on the solidus temperature of slag from the feed end to the middle of kiln. Figure 12 presents the effect of the replacement of CaO by MgO on the solidus temperature of slag from the middle of kiln to the discharge end. Figure 11 Effect of MgO on solidus temperature of slag from the feed end to the middle of kiln . 312 J. Univ. Sci. Technol. Beijing, Vol.12, No.4, Aug 2005 According to figures 9-12, the solidus temperature of slag increases with the CaO/SiO2 ratio increasing. The replacement of CaO by MgO also increases the solidus temperature of slag. Figure 12 Effect of MgO on solidus temperature of slag from the middle of kiln to discharge end. 4 Conclusions (1) The process model, combined thermodynamic modeling with heat transfer calculations, has been de- veloped to simulate the chemical reactions, mass and heat transfer and heat balance in the kiln. The injec- tion of air into the slag and the temperature profile along the kiln have been modeled. The effect of (CaO+MgO)/SiO2 on the solidus temperature of slag has also been modeled and discussed. Some optimized results have been presented in this paper. (2) The solidus temperature of slag increases with the CaO/SiO2 ratio increasing. The replacement of CaO by MgO also increases the solidus temperature of slag. References [1] S.E. James and C.O. Bounds, Recycling lead and cad- mium, as well as zinc from EAF dust. [in] T.S. Mackey and R.D. Prengaman ed. Lead-Zinc 90, The Minerals, Metals and Materials Society, 1990, p.477. [2] C.O. Bounds, and J.F. Pussaterl, EAF dust processing in the gas-fired flame reactor process, [in] T.S. Mackey and R.D. Prengaman ed. Lead-Zinc 90, The Minerals, Metals and Materials Society, 1990, p.511. [3] U. Itoh, Y. Matsushige, K. Watanbe, and Y. Yoneoka, Treatment of zinc leaching residue by rotary kiln, J. Min. Metall. Inst. Jpn., 98(1982), No.2, p.383. [4] Y. Kusuda and T. Imai, Treatment of zinc leaching residue by Waelz kiln, [in] H. Henein and T. Oki, ed. First Inter- national Conference on Processing Materials for Proper- ties, The Minerals, Metals and Materials Society, Warrendale, Pennsylvania, 1993, p.405. [5] Y. Yoneoka and Y. kusuda, Improvement of residue treatment operation by rotary kiln, J. Min. Metall. Inst. Jpn., 102(1986), No.4, p.261. [6] K. Mager, U. Meurer, et al., Recovery of zinc oxide from secondary raw materials: new developments of the Waelz process, [in] D.L. Stewan, R. Stephens, and J.C. Daley ed. Recycling of Metals and Engineered Materials, The Min- erals, Metals and Materials Society, 2000, p.329. [7] P.F. Tan and C.F. Zhang, Modeling of accessory element distribution in copper smelting process, Scand. J. Metall., 26(1997), No.3, p.115. [8] P.F. Tan and C.F. Zhang, Thermodynamic modeling for direct lead processes, [in] B. Mishra eds. EPD Congress 1998, San Antonio, Texas: The Minerals, Metals and Ma- terials Society (TMS), 1998, p.815. [9] P.F. Tan and Dieter Neuschuetz, A thermodynamic model of nickel smelting and direct high-grade nickel matte smelting processes: model development and validation, Metall. Mater. Trans. B, 32B(2001), p.341. [10] P.F. Tan and Dieter Neuschuetz, A thermodynamic model of nickel smelting and direct high-grade nickel matte smelting processes: distribution behaviors of Ni, Cu, Co, Fe, As, Sb and Bi, Metall. Mater. Trans. B, 32B (2001), p.353. [11] P.F. Tan, Inaki Hurtado, Dieter Neuschuetz, and Gunnar Eriksson, Thermodynamic modelling of PCDD/Fs forma- tion in thermal processes, Environ. Sci. Technol., 35(2001), No.9, p.1867. [12] P.F. Tan, Modeling and control of dioxin formation during iron ore sintering operations, VDI Verlag, Duesseldorf, Germany, 2002, ISBN: 3-18-365705-8. [13] P. Koukkari, I. Laukkanen, and S. Liukkonen, Combina- tion of overall reaction rate with Gibbs energy minimiza- tion, Fluid Phase Equilibria, 136(1997), p.345. [14] K. Penttilä, The Simulation Model of a TiO2-Calcination Kiln (in Finnish) [Dissertation], Helsinki University of Technology, 1996. [15] C.W. Bale, Chartrand P. Degterov S.A., G. Eriksson, K. Hack, Ben Mahfoud R., J. Melançon, A.D. Pelton, and S. Petersen, FactSage thermochemical software and data- bases, CALPHAD, 26(2002), p.189. [16] Eberhard Saage and Uwe Hasche, Optimization of the Waelz process at the B.U.S Zinkrecycling Freiberg GmbH, Erzmetall, 57(2004), No. 3, p.138.