# Modelling for Flash Calcination and Surface Area Development of Dispersed Limestone Particles.

код для вставкиСкачатьDev. Chem. Eng. Mineral Process., 8(3/4),pp.233-243, 2000. Modelling for Flash Calcination and Surface Area Development of Dispersed Limestone Particles 2. Ying, 2. Chuguang*, L. Zhaohui and S. Xuefeng National Laboratory of Coal Combustion, Huuzhong University of Science and Technology, Wuhun 430074, ?? R. CHINA ~ ~ ~ ~~ ~~ A mathematical model for calcination, sintering and surface area development of dispersed limestone parricles is presented in the papel: The model describes the calcination by considering heat transfel; mass transfer of CO1 through a porous CaO layer, and the chemical kinetics. Difision of CO, in a porous product layer is discussed in detail. The extent of calcination, BET surface area, temperature and pressure distribution in a particle were calculated. The extent of calcination and BET surface area, obtained from the simulation, are consistent with experimental data. Introduction The emission of sulfur dioxide from coal fired boilers is generally viewed as a major contributor to acid rain and as a consequence it is important to reduce the amount of SOl released. Dry sorbent injection into pulverized coal boilers is a simple technology with relatively low cost, and it is especially suitable for the retrofit of existing power plant. Dry sorbent injection into pulverized coal boilers occurs at high temperatures (11OO-15OO K). The sorbent initially calcinates before undergoing sulfation, the reactions can be expressed as follows. * Authorfor correspondence. ~ ~ 233 Z. Ying, Z Chuguang, L.Zhaohui and S Xuefeng CaCO,(s) + CaO(s)+ CO,(g) + CaCO, MgCO, (s) + CaO MgO(s) 2C0,(g) cuo(s)+sO,(g)++o,(g) + CuSO,(s) (3) Three potential rate-limiting processes are involved in the calcination process of limestone. (1) Heat transfer to the surface of the particle and then through the porous CaO product layer to the reaction interface. (2) Chemical reaction at the interface. (3) Mass transfer of the COz released at the reaction zone away from the interface through the product layer. Many investigators have studied experimentally the calcination and sintering processes of dispersed limestone particles, and they have developed some mathematical models to describe the processes. Borgwardt [1J found that the decomposition reaction was kinetically controlled except for the final stage of decomposition where the diffusion of C02 through the product layer was rate limiting. Therefore, he suggested the calcination rate was proportional to the BET surface area of unreacted CaC03. Silcox et al. [2] developed a calcination model where the diffusion of evolved COz was incorporated. One conclusion from this study was that the escape of C02 does not significantly slow down the calcination reaction rate for small particles. Keener and Kuang [3] proposed a structural pore-development model for calcination assuming that the particle temperature was uniform and the effect of sintering could be omitted. The model was correct at lower temperatures, however the difference between the model predictions and experimental data increases significantly at higher calcination temperatures. Milne et al. [43 proposed an experimentally modified shrinking core model after analyzing previous experimental data obtained from an entrained flow reactor. They proposed that the decomposition process occurred in a complex reaction zone instead of only a distinct reaction interface. A shrinking core model with a dependence on particle size to the 0.6 power was shown to best describe the experimental data. Hu and Scaroni [5] developed a model considering heat and mass transfer and chemical kinetics for calcination of pulverized limestone particles under furnace injection. It was sufficient to describe the calcination process, but the treatment of 234 Modelling for Flash Calcinationof Dispersed Limestone Particles sintering and C02 diffusion in the product layer needed further improvement. The objective of this study was to develop a model considering heat transfer, mass transfer and chemical reaction in order to better simulate the calcination and sintering processes. The modelling effort was focused on describing COz diffusion in the product layer in detail, and surface area development during calcination and sintering. Model Description I Thermal Decomposition The thermal decomposition of CaCO3 is a reversible reaction and has an equilibrium decomposition pressure of COz, which was established from the work of Hills as: Pe = 1.826~10~ exp(-19680/T) (4) The reaction rate is expressed as: Rate = -k, .Acnco3 ACoCO3 = S g 'Ncac03 mi'4cac03 The effect on the reaction rate of the partial pressure of the C02 gaseous product surrounding the solid is obvious, and there have been studies by different investigators [6,7]. However, general agreement has not yet been reached. We have used the same formula as Hu and Scaroni [5]: where kf' is the rate equation established by Borgwardt [I] in a calcination experiment on small dispersed limestone particles, and is as follows: k,' = 3 . 6lo7 ~ exp(-205000/ R T ) (8) Thus, the conversion fraction (x) of CaC03 to CaO can be expressed by: ln(1- x) = -k, .S, Mcaco3.t (9) 235 2.Ying, 2 Chuguang, L Utaohui and S Xuefeng I1 Heat 'kanfer For a spherical particle, the unsteady-state heat transfer equation is: Initial condition: T = To t=O (aT/ar) = o Boundary conditions: r = 0 r = ro where h is the convective heat transfer coefficient (W/m2 K): h=Nu*LId, (11) 1 1 Nu = 2.0 + 0.69Re' PrT (12) where k is the thermal conductivity of a limestoneparticle. For uncalcined limestone, k is 1.64 Wim K. For fresh calcines, k is 0.084 W/m K. For partially calcined limestone, k is calculated using the following equation: k = 1/((1- x)/k, + X/k2) (13) For fully calcined and partially sintered limestone: k = 0.69 - 6.07X S (14) I11 Mass Zhnsfer The unsteady-state C02 diffusion equation is: [ i a 7% r De $1 aP at = -- RTG, t=O Initial condition: Boundary conditions: r = 0 p=o aplar = 0 r = ro - De (ap/ar) = hd (P - P b where h,, is the convective mass transfer coefficient ( d s ) : h, = Sh*D l d , (16) 1 1 Sh = 2.0+0.6Re2Sc3 236 (17) Modelling for Flash Calcination of Dispersed Limestone Particles where D,is the effective diffusion coefficient (m2/s), including the ordinary diffusion given by: coefficient (DAB)and Knudsen diffusion coefficient (DK) Tt D, =435.7 (18) '-i D,=-d 3 2RT M C O 2 & D,= -(l/DM + 1/D, )-' z - and d is mean pore diameter: 2 = 4~/(S,p(l- &)) (21) N Sintering Model and Su$me Area Development German and Munir [8] developed a sintering model based on a sintering mechanism proposed for the isothermal sintering of AI2O3,ZnO, Fez03 and TiOz. The model is: ((So -S)/Soy =k, * t (22) Borgwardt [9]studied experimentally the sintering process of fresh CaO under H20 and COz atmospheres and obtained the empirical correlation of sintering rate constant k, and the exponent "I: In k, = 1.485+ 0.558 In pco, - 1 16601T (23) In r = 0.0034T + (In pco, - 1.948)/ 44.9 (24) When the fresh CaO is obtained at t= t,which has a BET surface area of So, CaO sinters at T for a certain shell, its BET surface area decreases to S"*at t = t, : A fundamental assumption of the sintering model is that A&is proportional to AS: E2 = El - (&I /s, XSl - sz ) and S=S,p(I-E) (26) (27) 237 Z.Ying, Z Chuguang, L. Zhaohui and S Xuefeng Therefore, the porosity of the shell (So) at t = t, is E"" Therefore, the surface area and porosity of the shell at t = t,, are s a oand E a o : k=l (29) Results and Discussion A. Model Verification Model predictions are compared with experimental data for different sorbents and particle size fractions in Figure 1. For the different sorbents and particle sizes calcinated, the model fits the experimental data well at all temperatures. As the model incorporates only fundamental physical and chemical processes during heating and calcination, it can be concluded that it can adequately describe the mechanism of calcination under experimental conditions. 238 Modelling for Flash Calcination of Dispersed Limestone Particles lOum 1198K 0 1198K 1Oum experimental data(reference 1) 6 3 m 1473K .-.-.-A 1473K 63um experimental data (reference 5 ) - - - - -91.17m o 0.1 0.2 0.3 0.4 0.5 0.6 1473K 1473K 91.17um t h i s study ClS Figure 1. Comparison between m d e l predictions and experimental data for the calcination of dispersed limestone particles. B. Calculation Results and LXwussion After the particle is injected into the boiler furnace its temperature increases rapidly. However, the particle temperature is lower than the gas temperature during decomposition. A large inner temperature gradient in the particle exists because calcination is an endothermic reaction and the thermal conductivity of porous CaO is relatively poor. Figure 2 shows particle surface temperature as a function of residence time and particle diameter for different particle diameters injected into 1473 K nitrogen. It is predicted that the heat transfer from entrained gas to particle surface is very rapid for small particles, and the surface temperature is nearly constant during the course of calcination. Actually, the temperature distribution is also uniform. So, it is clear that the decomposition rate is controlled by chemical kinetics. However, the surface temperature is much lower than the gas temperature during calcination for larger particle diameters (91.17 pm, 152.33 pm). Figure 3 shows the calculated temperature distribution as a function of residence time for 91.17 pm particles injected into 1473 K nitrogen. It is obvious that there is a large temperature gradient in particle. This temperature gradient will cause a large difference in the extent of calcination within the particle because the decomposition rate is exponentially dependent on temperature. 239 Z. Ying, Z Chuguang, L Zhaohui and S Xuefeng 1500 I 1400 t4 l300 H \ llo0 I200 no0 loo0 1000 900 1 , r / r O = O .4 ,zr/rO,=O.6 r/rO=O.8 . I r/rO=l 0 0 0.1 0.2 0.3 0.4 0 0.5 0.1 t/s Figure 2. Surface temperature as a function of residence timefor direrent size limestone particles injected into 1473 K nitrogen. 0.2 0.3 0.4 0.5 t/s Figure 3. Temperature distribution within the particle as a function of residence time for 91.17pn limestone particles injected into 1473 K nitrogen. CaCO3 and MgC03 decompose rapidly during limestone calcination at high temperatures, and a large amount of COz is released. Figure 4 shows the distribution of COzpartial pressure within the particle as a function of residence time for 9 1. I7pn particles injected into 1473 K nitrogen. In the initial stages of calcination, the extent of calcination is small, and the pore structure of the product layer does not develop completely. The porosity and C02 diffusion coefficient are thus small so that C02 cannot diffuse across the product layer into surrounding atmosphere in time. The pressure of COz within the particle increases. Meanwhile, the maximum C02pressure increases as the particle diameter increases. Then, the partial pressure of COz decreases slowly because the extent of calcination increases and porosity of the product layer increases during the calcination process. The temperature and pressure distribution in the particle will significiantly effect the rate of calcination of limestone, furthermore effecting the conversion rate of CaC03 to CaO. Figure 5 shows the extent of calcination within the particle as a function of residence time for 91.17 pm limestone particles injected into 1473 K nitrogen. It is predicted that there exists a big difference between the particle periphery and center in the extent of calcination. The extent of calcination at the particle’s center is lower than that at the particle’s periphery. Therefore, a model that omits heat and mass transfer within the particle will not predict the calcination process of larger limestone particles adequately. 240 Modelling for Flash Calcination of Dispersed Limestone Particles 3r/rO=O.6 0.6 ------c---r/ro=o. 8 r/rO=O 0.4 0.2 2 0 0 0.1 0.2 0.3 0.2 0 0.4 Figure 4. CO,partial pressure distribution within the partkle as a function of residence timefor 91.17 pm limestone particles injected into 1473 K nitrogen. tn 0.4 0.6 t/s t/s Figure 5. Extent of calcination as a function of residence timefor 91.17 pm limestone particles injected into 1473 K nitrogen. 30 13.65um Y 0 0 0.2 0.4 t/s 0.6 Figure 6. BET surface area as a function of residence time for 13.65 pm limestone particle. 0 0.2 0.4 0.6 t/s Figure 7. BET surface area as a function of residence time for limestone particle of different diameters (T=1373K). In Borgwardt’s sintering experiment [9],the BET surface area of fresh CaO can be as high as 104 m2/g. Unfortunately, the fresh CaO looses most of its surface area due to high temperature sintering. Figure 6 shows BET surface area as a function of residence time for 13.65 j.un limestone particles calcined at different temperatures. It predicts that the maximum BET surface area is considerably lower than 104 m2/g, and does not occur at the maximum extent of calcination. Figure 7 shows BET surface area as a function of residence time for limestone particles of different diameter calcined at 1373 K. It predicts that the maximum BET surface area decreases as the particle diameter decreases, and it will take longer for it to reach the maximum BET surface area. The BET surface area is clearly correlated with calcination rate. 241 Z. Ying, Z Chuguang, L Zhaohui and S Xuefeng Conclusions A model including heat transfer, mass transfer and chemical kinetics is developed in this paper for limestone calcination and particle surface area development. For calcination of dispersed limestone particles, these processes all exert a significant influence on the calcination rate especially for larger limestone particles. There are large temperahue and COz partial pressure gradients within the particle. Therefore the extent of calcination in the particle is also significantly different, and a shrinking core model and other simple models cannot adequately simulate the dispersed limestone calcination. The BET surface area is closely correlated with the conversion rate. It is significantly lower than the BET surface area of the fresh CaO. Nomenclature b C 0 3 k/? k hfCaC03 R Rate s, sm T T X Surface area of CaC03 Mass diffksivity D dP Particle diameter Decompositionrate constants Molecular weight of CaCO3 Gas constant Decompositionrate Specific BET surface area of CaC03 BET surface area Time Temperature Extent of calcination Greek letters h Thermal conductivity of N2 z Tortuosity factor & Porosity P* True density References 1. Borgwatd~R.H. Calculation hetics and surface area of dispersed limestone particles. AIChE J., 2. Silcox, G.D.,Kramlich, J.C.,and Pemhing, D.W. Mathematkd model for the flash calcination of dispersed CaCO, and Ca(Om particles. lnd. Eng. Chem. Res., 1989.28: 155-160. 1985.31: 103-111. 242 Modellingfor Flash Calcination of Dispersed Limestone Particles 3. 4. Keener, S., and Kuang, S.J. Structural pore development model for calcination. Chem. Eng. COIXUIIU~.. 1992,117: 279-291. M h e , C.R., Silcox, G.D., and Pershing. D.W. Calcination and sintering models for application to high-temperature, short-time sulfation of calcium-based sorbents. Ind. Big. Chem. Res., 1990, 29: 139-149. 5. 6. 7. 8. Hu, N., and Scaroni, A.W. Calcination of pulvdzed limestone particles under fumace injection conditions. Fuel, 19%. 75: 177-186. Wang, Y., and Thoms0n.W.J. The effect of steam and carbon dioxide on calcite decomposition using dynamic X-raydiffraction. Chem. Eng. Sci., 1995.50: 1373-1382. Khinast, J., Knunma, G.F., Brunner, C., and Staudinger, G. Decomposition of limestone the influence of C& and particle size on the reaction rate. Chem. Eng. Sci., 1996.51: 623-634. German, R.M., and Munir, Z.A. Surface area reduction during isothermal sintaing. J. Am. Ceram. SOC.,1976.59: 379-383. 9. Borgwardt, R.H. Calcium oxide sinering in atmospheres containing water and carbon dioxide. Ind. Eng. Chem. Res., 1989.28: 493-500. 243 244

1/--страниц