Modeling Of Coal Gasification In An Internally Circulating Fluidized Bed Reactor With Draught Tube

Fuel 79 (2000) 69–77 www.elsevier.com/locate/fuel

Modeling of coal gasification in an internally circulating fluidized bed reactor with draught tube Y.J. Kim, J.M. Lee 1, S.D. Kim* Department of Chemical Engineering and Energy & Environment Research Center, Korea Advanced Institute of Science and Technology, Taejon 305-701, South Korea Accepted 14 June 1999

Abstract A predictive mathematical model is proposed based on the bed hydrodynamics, reaction kinetics and the empirical correlation of pyrolysis yields to predict gasification characteristics in an internally circulating fluidized bed gasifier with a draught tube. With the justifiable assumptions, steady state mathematical equations are derived and solved numerically. The simulated results of product gas composition, gas yield, carbon conversion, cold gas efficiency and calorific value of the product gas in each reaction region are compared with the obtained experimental data. The proposed model can explain the reaction behavior in the present reactor system within the range of variables studied. q 1999 Elsevier Science Ltd. All rights reserved. Keywords: Modeling; Gasification; Internally circulating fluidized bed; Draught tube

1. Introduction Gasification in a fluidized bed can be utilized to convert coal [1–7], biomass [8] and waste materials [8,9] into fuel/ synthesis gas. The intrinsic problems of coal conversion to heat or fuel/synthesis gases in fluidized beds are: high carbon losses due to coal shattering, and subsequent elutriation of fines, and low conversion of reactant gases due to gas bypassing [3,7,10–13]. To solve these problems a draught tube was inserted in a fluidized bed to divide the fluidized bed into two reaction zones [10–16]. Two reaction zones can be attained for coal combustion with air feeding in the draught tube and coal gasification with steam feeding in the annulus zone [10–15], or vice versa [16]. By fluidizing solids in the draught tube at a velocity about 7–10 times of minimum fluidizing velocity (Umf) and the annulus at 0:7–1:5Umf , it is possible to induce a gross circulation of bed material up the draught and down to the annulus region. Solid circulation within the reactor provides energy transfer from the combustion zone into the gasification zone and significantly reduces elutriation of fine coal particles from the reactor [17]. The longer residence time of fine char * Corresponding author. Tel.: 1 82-42-869-3913; fax: 1 82-42-8693910. E-mail address: [email protected] (S.D. Kim) 1 Present address: Power Generation Research Laboratory, Korea Electric Power Research Institute, Taejon 305-380, South Korea.

particles in the annulus region may provide much higher conversion level compared to a conventional fluidized bed gasifier. By installing a gas separator over the draught tube, high-calorific value gas in the annulus (gasification) zone can be obtained. To increase the calorific value of the product gas in the annulus zone, a draught tube having orifices at the bottom part was devised based on the findings in previous studies [12–18]. The internally circulating fluidized bed (ICFB) is similar in many ways to the traditional circulating fluidized bed (CFB). The main differences are in the modes of solid circulation and fluidization. Various reactor models have been proposed for conventional CFB but models for ICFB are quite sparse. To develop good reactor models, the bed hydrodynamics and the reaction kinetics are needed to predict reactor performance. Several CFB modeling approaches lead to the models with different degrees of sophistication [19]. While cyclone, downer or annulus in CFB are commonly treated as the well mixed or plug flow systems; modeling for the riser regions varies a lot from homogeneous [20] to heterogeneous modeling [21–26], from single region [20] to multiple region modeling [21– 23] and from zero- to three-dimensional modeling [24–26]. The existing models cannot be directly applied to the internally circulating fluidized bed gasifier (ICFBG) due to the different flow behavior in the riser (draught tube). A predictive model is developed to describe and characterize the performance of ICFBG with all the key

0016-2361/00/$ - see front matter q 1999 Elsevier Science Ltd. All rights reserved. PII: S0016-236 1(99)00128-3

70

Y.J. Kim et al. / Fuel 79 (2000) 69–77 Table 1 Analyses of Australian coal and char preparation

Fig. 1. Schematic diagram of the internally circulating fluidized bed gasifier. 1. flow meter, 2. steam generator, 3. orifice meter, 4. air plenum, 5. distributor, 6. overflow drain, 7. viewport, 8. draft tube, 9. main body, 10. separator, 11. freeboard, 12. screw feeder, 13. coal hopper, 14. cyclone, 15. condenser, 16. collector, 17. dust filter, 18. condenser, 19. gas sampling bottle, 20. I.D. fan.

operating parameters such as reaction temperature (780– 9008C), oxygen/coal ratio (0.30–0.53), coal-feeding rate (5.3–12.1 kg h 21), and steam/coal ratio (0.30–0.81). The proposed model is based on the bed hydrodynamics [18], reaction kinetics for combustion and gasification of coal [5] as well as the empirical correlation of pyrolysis yields [5]. In the proposed model, the ICFB is divided into two parts, namely the draught tube region being considered as a well mixed fluidized bed reactor and the annulus region being considered as a plug flow moving bed reactor.

Proximate analysis

Wt%, db

Ultimate analysis

Wt%, daf

Volatile matter Fixed carbon Ash Heating value (cal g 21)

26.37 63.91 9.72 6273

C H O N S

68.10 4.58 25.38 1.62 0.32

caps was used for steam supply into the annulus section. Four evenly spaced 30 mm diameter holes were drilled on the walls of the draught tube 40 mm above its bottom for solids circulation. A gas separator was installed to separate the gas streams emanating from the annulus and draught tube zones. To heat the reactor to ignition temperature of the coal (ù5008C) an electric heater (16 kW) was installed at the main reactor wall. The reactor was insulated by Kaowool to prevent heat loss through the reactor wall. An ash-drain port was installed at the bottom of the reactor and an overflow drain port was mounted 1.1 m above the distributor. The freeboard (0.45 m i.d.) section was expanded to reduced particle entrainment from the reactor. The coal was fed into the top of the reactor through a screw feeder, which was connected, to a coal hopper and a variable DC motor controller regulated the coal-feeding rate. Two cyclones (0.08 m i.d. and 0.32 m high) were installed at the outlet of the reactor. The product gas was cooled through a condenser and entrained fines in the product gas were collected in a bag filter. The gas sampling probes were mounted at the outlet of the condenser. At the beginning of the experiment, only air was fed into the reactor until the bed temperature reached 450–5008C by the electric heating. Thereafter, the electric heater was turned off and coal was fed into the gasifier. When the desired reaction temperature was reached, steam was introduced into the gasifier. When the gasifer operation reached steady state, the product gases from the draught tube and annulus zones were sampled, and the amount of collected particles in the cyclone was measured. The product gas was analyzed by using gas chromatography (HP 5890 series II). Particles size of the bed material (sand) was 390 mm and the static bed height was 0.8 m from the distributor. The proximate and ultimate analyses of the coal (defaultdp: 1–5 mm) are given in Table 1.

2. Experimental Experiments were carried out in a stainless steel column (0.3 m i.d. and 2.7 m high) with a centrally located draught tube (0.1 m i.d. and 0.9 m high) as shown in Fig. 1. The air plenum was divided into two parts to supply air into the draught tube and steam into the annulus sections separately. For air supply to the draught tube, a distributor (0.1 m i.d.) with seven bubble caps …4 holes × 2:5 mm i:d: in each bubble cap was used. Also, a conical plate having inclined angle of 608 relative to the horizontal plane with 18 bubble

default3. Modeling of ICFBG To formulate a mathematical model, the ICFB is divided into two parts, namely the draught tube region being considered as a well mixed fluidized bed reactor and the annulus region being considered as a plug flow moving bed reactor. In the annulus region, the bed material (sand) and char move downward under gravity flow and the gases (steam and

Y.J. Kim et al. / Fuel 79 (2000) 69–77

71

Table 2 Kinetic parameters for the reactions Reaction

Combustion Gasification H2 1 1=2O2 ! H2 O CO 1 1=2O2 ! CO2

Temperature range (8C) 500–575 575–700 700–800 700–850 – –

4 21

21

7:58 × 10 s atm 0.44 s 21 atm 21 0.045 s 21 atm 21 6:47 × 103 s21 atm21 3:09 × 1011 m3 mol21 s21 8:83 × 1011 m3 mol21 s21

bypassed air) flow upward [27]. The following assumptions are made for the modeling of the ICFBG system. 1. The gasifier is operating under steady state and isothermal conditions. 2. Since the annulus zone (gasification) occupies almost the entire reactor volume, it can be assumed that the feeding coal from the top of the reactor enters the annulus region only and the particles are withdrawn from the overflow in the annulus region. 3. The hydrodynamics of the draught tube region (fluidized bed) are described by the two-phase theory of fluidization [3,6,28–30]. The bubble phase is plug flow without particles and emulsion phase is completely mixed flow at the incipient fluidization condition. 4. Drying and devolatilization of the particles take place in the freeboard regions of ICFBG and the products of volatiles are distributed uniformly between the annulus and draught tube regions. The obtained char contains only carbon and ash. 5. In the freeboard region, homogeneous reactions occur and the gas phase is free of solids in plug flow. 6. At the distributor level …z ˆ 0†, gas bypassing occurs from the draught tube to the annulus zone and vice versa, thus gas concentrations of each reaction zone after gas bypassing are the inlet concentrations of each zone. 3.1. Combustion and gasification The overall coal gasification involves pyrolysis and the homogeneous reactions in the freeboard region, and the heterogeneous reactions in the main bed region. The following heterogeneous reactions are assumed to occur between solid carbon and gaseous reactants [5]: Combustion reactions [5,27,29,31] C 1 aO2 ! 2…1 2 a†CO 1 …2a 2 1†CO2 :

E (kJ mol 21)

k0 ()

…1†

Steam gasification reactions [5,32] C 1 H2 O ! CO 1 H2

…2†

CO 1 H2 O ! CO2 1 H2

…3†

C 1 2H2 O ! CO2 1 2H2

…4†

113 27.6 9.6 167 99.8 99.8

C 1 bH2 O ! …2 2 b†CO 1 …b 2 1†CO2 1 bH2

Reference

Lee et al. [5] Lee et al. [5] Haslam [35] Tesner [36]

…5†

Here a is a system constant, dependent on the reaction conditions, which determines the primary product distribution of CO in the combustion products [27]. The molar ratio of CO/CO2 in the combustion reaction follows the Arrhenius temperature relation. There are different expressions to calculate the distribution of combustion products [33,34]. According to Arthur [33], for the atmospheric beds, this is given by   6234 ; …6† CO=CO2 ˆ 2400 exp 2 Ts where Ts (surface temperature of the char particle) is assumed equal to the bed temperature. The a can be determined as a function of temperature from Eq. (6). Values of a at 1023 and 1173 K are 0.58 and 0.54, respectively. In the given experimental temperature range, a does not produce any significant effect on the model prediction. In Eq. (5), …2 2 b†=b represents the fraction of steam consumed by the reaction path (2) and 2…b 2 1†=b represents the fraction of steam consumed by the reaction path (4). Matsui et al. [32] experimentally determined b that decreases with increasing temperature in the range of 1.1– 1.5 at 750–9008C. In this study, b value from Matsui et al. [32] was adopted for the modeling of coal gasification. In a previous study [5], the rate equations for combustion and steam gasification are determined based on the shrinking-core model as dX ˆ kpnO2 dt

or H2 O …1

2 X†2=3 ;

…7†

where X and pO2 or H2 O are carbon conversion and partial pressure of O2 or H2O, respectively. The activation energies and the reaction rate constants for the combustion and steam gasification reactions are determined from an Arrhenius plot with the data obtained from the thermobalance reactor. In addition to the above heterogeneous reactions, gas phase combustion reactions of H2 and CO may occur. H2 1 1=2O2 ! H2 O

r3 ˆ k3 CH2 2 CO2 =CCO

…8†

CO 1 1=2O2 ! CO2

r4 ˆ k4 CCO CO2

…9†

where ri and ki are the reaction rate and the reaction rate constant of reaction i, respectively. The overall reaction

72

Y.J. Kim et al. / Fuel 79 (2000) 69–77 1 0 0 Fi;d ˆ Fi;a fad 1 Fi;d …1 2 fda †;

…14†

Va1 ˆ Va0 …1 2 fad † 1 Vd0 fda ;

…15†

Vd1 ˆ Va0 fad 1 Vd0 …1 2 fda †;

…16†

fad ˆ 5:06…Ud0 =Umf †0:885 …Ua0 =Umf †0:082 ;

…17†

3.2. Coal pyrolysis

fda ˆ 2:18…Ud0 =Umf †20:467 …Ua0 =Umf †20:048 :

…18†

Lee et al. [5] measured the product gas yield from pyrolysis in a fluidized bed reactor using the same coal in this study and compared the results with the empirical correlations of Ma et al. [29] and Loison and Chauvin [38]. They found that their correlations [29,38] do not predict our pyrolysis data accurately. Therefore, the following assumptions are made for modeling of ICFBG to determine the overall reactor performance since the amount of gas produced from coal pyrolysis contributes greatly to the total gas production from the ICFBG. Drying and devolatilization of coal occur instantaneously in the freeboard region of ICFB. The obtained product gas yields from pyrolysis in a fluidized bed reactor are given by [5]:

Here, C, F, V and U represent the concentration, molar feed rate, volumetric flow rate of gas and gas velocity, respectively. The subscripts i, a and d denote gas components (H2O, O2, N2, H2, CO and CO2), the annulus region and the draught tube region while the superscripts 0 and 1 denote the inlet condition and boundary condition at the bottom of the reactor. At the top of the annulus region …z ˆ H†, the carbon balance can be written as follows: The sum of carbon feed rate from the draught tube to the annulus and that from the coal hopper equals the sum of carbon feed rate in the downward moving bed (annulus region) and carbon discharge from the reactor.

yH2 ˆ 8:233 × 1025 T 2 0:073;

Wfc …1 2 Xd † 1 Fc fchar ˆ Wfc …1 2 XaH † 1 Fc fchar …1 2 XaH †; …19†

orders of the homogeneous are second order [28,37] and the rate constant (k) can be represented by the Arrhenius expression. Weimer and Clough [28] and Ma et al. [29] who considered homogeneous combustion reactions in their models used the above rate expression. The values of the frequency factor (k0) and activation energy (E) of each reaction used in the model prediction are deduced from the literature as shown in Table 2.

yCO ˆ 2:168 × 1024 T 2 0:102; 26

yCO2 ˆ 2:999 × 10 T 1 0:039;

…10†

yCH4 ˆ 2:561 × 1025 T 1 0:011 where T is in Kelvin and yi (kg/h) is the gas yield to coalfeed ratio (kg/h) from pyrolysis on a dry ash-free basis. The distribution of the pyrolysis products affects the compositions of gas produced in each region of the gasifier. Thus, in the present model, fraction of pyrolysis in the draught region can be assumed to be 0.5 from the experimental data. 3.3. Boundary conditions In the ICFBG, the following boundary conditions can be made: At the distributor level (z ˆ 0), gas concentrations of each reaction zones can be obtained from assumption (6). The fraction of gas bypassing from the draught tube to the annulus (fda) and from the annulus to the draught tube (fad) have been correlated from the data of Ahn [18]. In the moving bed (annulus region) 1 1 ˆ Fi;a =Va1 Ci;a

i ˆ H2 O; O2 ; N2 ; H2 ; CO; CO2

…11†

where W, fc, Fc, fchar and XaH are mass flow rate of circulating solids, weight fraction of carbon in the bed, coal feed rate, weight fraction of char in the coal and carbon conversion at the top of the annulus region, respectively. 3.4. In the annulus region (moving bed) Mathematical models for a moving bed gasifier have been reported in the literature [27,31,39–41]. In a moving bed gasifier, the coal moves downward under gravity flow and a mixture of steam and bypassed air flow upward countercurrently. Since the annulus moving bed is assumed as a pseudoplug flow reactor, the steady state mass balances for fixed carbon, reactant and product gases can be written as For carbon: dX ˆ {k1 pnO2 1 k2 pnH2 O }…1 2 X†2=3 : dt

…20†

For reactant and product gases RO2 ˆ 2dCO2 =dt ˆ k1 PnO2 …1 2 X†2=3 …rs fc =MB †a 1 0:5r3 1 0:5r4 ;

…21†

In the fluidized bed (draught tube region) 1 1 ˆ Fi;d =Vd1 Ci;d

i ˆ H2 O; O2 ; N2 ; H2 ; CO; CO2 ;

…12†

RH2 O ˆ 2dCH2 O =dt ˆ k2 PnH2 O …1 2 X†2=3 …rs fc =MB †b 2 r3 ; …22†

where 1 0 0 ˆ Fi;a …1 2 fad † 1 Fi;d fda ; Fi;a

…13†

RH2 ˆ 2dCH2 =dt ˆ 2RH2 O ;

…23†

Y.J. Kim et al. / Fuel 79 (2000) 69–77

73

Table 3 Basic equations in the draught tube region (fluidized bed) 1. Two phase model parameters (a) Minimum fluidizing velocity [42] …1mf ˆ 0:5; 1m ˆ 0:4; rs ˆ 2:4† Umf ˆ …m=rg dp †{‰28:72 1 0:0494dp3 rg …rs 2 rg †g=m2 Š1=2 2 28:7} p (b) Bubble volume fraction [43] d ˆ 1 2 ‰1 1 ……U 2 Umf †=0:35 gDt †Š21 (c) Bubble-rise velocity Ub ˆ …Ud 2 2…1 2 d†Umf †=Umf (d) Interchange coefficient between the bubble and emulsion phases [44] Kbe ˆ 11=db 2. Mass-balance equation (a) Gas phase reaction rate equations in emulsion phase RO2 ˆ 2dCO2 =dt ˆ k1 POn 2 …1 2 X†2=3 …rs fc =MB †a 1 0:5r3 1 0:5r4 RH2 O ˆ 2dC H2 O =dt ˆ k2 PHn 2 O …1 2 X†2=3 …rs fc =MB †b 2 r3 RH2 ˆ 2dCH2 =dt ˆ 2RH2 O RCO ˆ 2dCCO =dt ˆ 2{2…1 2 a†k1 POn 2 1 …2 2 b†k2 PHn 2 O }…1 2 X†2=3 …rs fc =MB † 1 r4 RCO2 ˆ 2dCCO2 =dt ˆ 2{…2a 2 1†k1 POn 2 1 …b 2 1†k2 PHn 2 O }…1 2 X†2=3 …rs fc =MB † 2 r4 (b) Gas phase reaction rate equations in bubble phase RO2 ˆ 2dCO2 =dt ˆ 0:5r3 1 0:5r4 RH2 O ˆ 2dC H2 O =dt ˆ 2r3 RH2 ˆ 2dCH2 =dt ˆ 2RH2 O RCO ˆ 2dCCO =dt ˆ r4 RCO2 ˆ 2dCCO2 =dt ˆ 2r4 (c) Reactant gas mass balances …i ˆ O2 ; H2 O; H2 ; CO; CO2 ; N2 † 2Ub dCi;b =dz ˆ Kbe …Ci;b 2 Ci;e † 1 Ri;b 2…1 2 d†Umf dCi;e =dz ˆ …1 2 d†…1 2 1mf †Ri;e 2 dKbe …Ci;b 2 Ci;e †

RCO ˆ 2dCCO =dt ˆ 2{2…1 2 a†k1 PnO2 1 …2 2 b†k2 PnH2 O } × …1 2 X†2=3 …rs fc =MB † 1 r4 ;

…24†

RCO2 ˆ 2dCCO2 =dt ˆ 2{…2a 2 1†k1 PnO2 1 …b 2 1†k2 PnH2 O } × …1 2 X† …rs fc =MB † 2 r4 ;

…25†

2=3

where r s and MB are solid density and molecular weight of carbon, respectively. The above mass balance equations can be rearranged to eliminate dt (dt ˆ dz=us or dt ˆ dz=Ua in solid or gas phase) as follows. For carbon dX 1 ˆ {k pn 1 k2 pnH2 O }…1 2 X†2=3 : dz us 1 O2

…26†

For reactant and product gases 2Ua dCi =dz ˆ Ri

i ˆ O2 ; H2 O; H2 ; CO; CO2 ; N2 ;

…27†

where z is the distance from the bottom of the reaction zone and us is the solid falling velocity that is proportional to the solid circulation rate (Ws) since particles are assumed to have same size. According to the cold model test [18], the solid circulation rate has been correlated by the following equation: Ws ˆ 2:63 × 1025 rs …1 2 1mf †   Umf

Ua × Ud Umf

1:19

dor dp

!1:29 

H Hs

0:873

;

…28†

where 1 mf, dor, H and Hs are void fraction at minimum fluidizing condition, orifice diameter, bed height at the

fluidization condition and static bed height, respectively. In steady state, solid falling velocity (us) is obtained from the solid circulation rate and bulk density of solid (r bulk) in the annulus region as us ˆ

Ws : rbulk

…29†

From the above differential equations and boundary conditions, carbon conversion and product gas compositions in the annulus region are calculated as a function of the axial height of the reactor with the assumed carbon conversion at the bottom of the annulus region …Xa1 †. Carbon conversion in the draught tube region (Xd) is calculated from the boundary condition at the top of the annulus region. 3.5. In the draught tube region (fluidized bed) The basic equations in the draught tube region are summarized in Table 3 where 1 m, m , r s, g, d , Dt, Kbe and db are void fraction in a fixed bed, gas viscosity, gas density, gravitational constant, bubble-volume fraction, bed diameter, gas interchange coefficient between bubble and emulsion phase and bubble diameter, respectively. The subscripts b and e denote the bubble and the emulsion phase. From the calculated carbon conversion in the draught tube region (Xd), gas concentration profile along the height is calculated from the hydrodynamic parameters, mass balance and gas reactions. From the carbon conversion profile in the annulus and the draught tube regions, the average carbon conversion in the bed (Xt) can be formulated as Xt ˆ

Ad …1 2 d†…1 2 1mf †HXd 1 Aa …1 2 1m †Hs X a ; Ad …1 2 d†…1 2 1mf †H 1 Aa …1 2 1m †Hs

…30†

74

Y.J. Kim et al. / Fuel 79 (2000) 69–77

conversion is the bed, Xt, is calculated from Eq. (30). (6) From the gas concentration profile along the bed height, the average carbon conversion, X 0t , is calculated. (7) An iteration process is repeated until the average carbon conversion …X 0t † converges to Xt. (8) With the calculated carbon conversion …X 0t †, the product gas compositions, gas concentration profile and the reactant gas conversion of each reaction region can be predicted from the pyrolysis data. (9) Product gas compositions, the reactant gas conversion and gas yields from each region at the reactor exit can be obtained after completion of homogeneous reaction in the freeboard region in the bed.

4. Results and discussion Fig. 2. Effect of reaction temperature on compositions of the product gas in the annulus region.

where A and X a are cross-sectional area and the average carbon conversion in the annulus region, respectively. 3.6. Calculation procedure of the model The calculation procedure of the present model is as follows: (1) Assume the solid conversion in the bottom of the annulus region, Xa1 . (2) Carbon conversion and gas concentration profiles along the bed height are calculated from the mass balance equation in the annulus region. (3) Carbon conversion in the draught tube region (Xd) is calculated from the boundary condition at the top of the annulus region. (4) The bed hydrodynamic parameters are determined from the input conditions and the gas concentration profile along the bed height in the draught tube region is calculated from the hydrodynamic parameters, mass balance and gas reaction at a given Xd. (5) Average carbon

Fig. 3. Comparison of gas compositions from pyrolysis and gasification with variation of reaction temperature. The solid and dotted lines represent the model prediction and pyrolysis, respectively.

4.1. Effect of reaction temperature The reaction temperature is expected to be one of the most important operating variables affecting the performance of coal gasifier since the main gasification reactions are endothermic. Hence the reaction temperature should be at the highest tolerable level, the materials construction, ash fusion and production of undesirable gases such as NOx will limit the reaction temperature [12,13,45]. In the present study, the construction material of the gasifier restricts the maximum operable temperature to 9008C. To determine the effect of reaction temperature on gasifier performance, the other experimental variables such as gas velocities to the draught tube …Ud ˆ 10Umf † and the annulus …Ua ˆ 1:4Umf † were held at the constant values. Thus, residence time of the reactant gases, solids circulation rate and gas-bypassing properties are very similar in the both regions. Though a detailed material balance could not be established due to the difficulty of estimating the tar yield accurately, the carbon balance can be determined from a previous study [12]. The effect of reaction temperature on the product gas composition in the annulus region is shown in Fig. 2 where solid lines represent the model prediction. The compositions of the product gas in the annulus region are H2 (44.9–49.7%), CO (23.6–28.1%), CO2 (14.8–18.9%), CH4(7.3–12.6%). The concentrations of H2 and CO increase with increasing reaction temperature [4,45,46] due to the endothermic gasification reaction of steam-char and pyrolysis, whereas the concentrations of other gases decrease. The gas compositions obtained in this study are comparable to that of Judd [10] who used a gap height type draught tube at higher reaction temperatures than that in the present study. Therefore it can be claimed that the draught tube with orifices is superior to the gap height type for producing medium-CV gas in the annulus region. As can be seen in Fig. 2, our model predicts the experimental data quite well. The gas yields from coal pyrolysis and gasification at different reaction temperature are shown in Fig. 3 where the solid and the dotted lines represent the values from the

Y.J. Kim et al. / Fuel 79 (2000) 69–77

Fig. 4. Comparison of calorific value of the product gas with reaction temperature. The solid and dotted lines represent the model prediction and linear regression, respectively.

model and the correlation for pyrolysis, respectively. The slopes and yields of H2, CO and CO2 from gasification are higher than those from pyrolysis due to steam-char and oxygen-char reactions, whereas yields of CH4 agree well between coal gasification and pyrolysis. Also, the slopes of H2 and CO yields are steeper than that of CO2 in gasification owing to the rapid increase of gasification reaction. As can be seen in Table 2, the rates of combustion and gasification, respectively, at 9008C are 1.1 and 7.0 times faster than those at 7808C. Compositions of the product gas in the combustion zone (draught tube) exhibit the same trend as that in the gasification zone (annulus). The increase of H2 and CO yields may come from volatile mater release from pyrolysis and high steam bypassing from the annulus to the draught tube region [12]. The model predicts the

Fig. 5. Effect of reaction temperature on gas yield, carbon conversion and cold gas efficiency of the product gas.

75

experimental data fairly well except the slight overprediction of H2 yield. The effect of reaction temperature on the calorific values of the product gas in the present and previous studies [11,47,48] is shown in Fig. 4. Calorific value of the product gas from the annulus region is 11–12 MJ m 23 that is much higher than that in the annulus zone of the ICFB with a gap height type draught tube [11]. In the annulus zone, the calorific value decreases with increasing temperature due to the reduction of hydrocarbons and CH4 yields [12]. Since hydrocarbon yields were not accounted for in the model, the model under-predicts the calorific value of the product gas and increases slightly due to an increase in gasification reaction. In the draught tube region, calorific value of the product gas is 3.3–4.7 MJ m 23, which is comparable to the conventional fluidized bed [47] or the spout bed gasifier [48]. Therefore, it can be claimed that the reactor performance with the orifice type draught tube is superior compared to the gap height type draught tube in producing medium-calorific value gas in the annulus region due to the reduction of gas bypassing from the draught tube to annulus regions [12]. The effects of reaction temperature on gas yield, carbon conversion and cold gas efficiency of the product gas are shown in Fig. 5. Though carbon conversion caused by pyrolysis contributes large portion of the total carbon conversion, the effect of pyrolysis on the total carbon conversion decreases with increasing reaction temperature. Gas yield, carbon conversion and cold gas efficiency increase in the both regions due to the increase of reaction rates for steamchar, oxygen-char and pyrolysis with increasing reaction temperatures. The model predicts well the experimental data in the both regions. 4.2. Effect of oxygen/coal ratio To determine the effect of oxygen/coal mass ratio on the gasifier performance, reaction temperature, gas velocities to the draught tube …Ud ˆ 10Umf † and annulus …Ua ˆ 1:4Umf † were held at constant values. Oxygen/coal mass ratio is controlled by air and oxygen flow rates supplied into the draught tube. The effect of oxygen/coal mass ratio supplied to the draught tube on the product gas composition in both regions at a given steam/coal ratio, together with the predicted values from the model, is shown in Fig. 6. In the annulus region, the gas contents (vol.%) of H2, CO and CH4, and calorific value of the product gas decrease, but CO2 content increases due to the increase of oxygen content in the bypassing gas from draught tube to annulus region with increasing O2/coal ratio in the draught tube [12]. As in the annulus region, compositions of the product gas in the draught tube region exhibit the same trend. However, the effect of O2/coal ratio in the draught tube region is more pronounced compared to the annular region due to the higher oxygen partial pressure.

76

Y.J. Kim et al. / Fuel 79 (2000) 69–77

Fig. 6. Effect of O2/coal (kg kg 21) ratio on the compositions of the product gas in the annulus region.

Fig. 8. Effect of coal feed rate on compositions of the product gas in the annulus region.

The effect of oxygen/coal mass ratio on the calorific value of the product gas in both regions at a given steam/coal ratio, together with the predicted values from the model, is shown in Fig. 7. Since combustion of CH4 is not considered in the present model, the model over-predicts composition of CH4 (Fig. 6) and the calorific value of the product gas in the annulus region is 10.7 MJ m 23 despite the decrease in H2 and CO concentrations with increasing O2/coal ratio. Whereas, calorific value of the product gas in the draught tube region decreases slightly due to the significant decrease in H2 and CO concentration and low concentration of CH4 with increasing O2/coal ratio.

oxygen ratio, together with the model prediction, is shown in Fig. 8. As can be seen, the model data agrees well with the experimental one and the concentrations of H2, CO and CH4 increase with increasing coal feed rate due to an increase in pyrolysis yields, whereas that of CO2 decreases [1,2,12,45,46,48]. Thus, the calorific value of the product gas in the annulus region increases from 8.56 to 13.22 MJ m 23 due to an increase in combustible gas concentrations. Since the oxygen/coal ratio decreases with increasing coal feed rate, product gas yield and carbon conversion decrease [12,46]. Cold gas efficiency does not exhibit any noticeable variation.

4.3. Effect of coal feed rate

4.4. Effect of steam/coal ratio

The effect of coal feed rate on compositions of the product gas in the annulus region at a constant steam/

The effect of steam/coal mass ratio supplied to the annulus region on compositions of the product gas at a coal feed rate of 7.56 kg h 21 and a O2/coal mass ratio of 0.30, together

Fig. 7. Comparison of calorific value of the product gas with O2/coal (kg kg 21) ratio. The solid and dotted lines represent the model prediction and linear regression, respectively.

Fig. 9. Effect of steam/coal (kg kg 21) ratio on the gas compositions of the product gas in the annulus region.

Y.J. Kim et al. / Fuel 79 (2000) 69–77

with model prediction, is shown in Fig. 9. The effect of steam/coal ratio on the gasifier performance is insignificant since the reaction rate of steam-char is not sensitive to the steam/coal ratio [12,45,46]. The model prediction does not show any variation in the product gas composition as found in the experimental data. The calorific value, gas yield, carbon conversion and cold gas efficiency of the product gas remain nearly constant. Since the proposed model predicts well the experimental data, it can be claimed that the presented model can be utilized to predict the performance of ICFBG. 5. Conclusions Coal gasification was carried out in an ICFB with an orifice type draught tube over a temperature range of 1053–1173 K at atmospheric pressure. In the ICFBG, low-calorific value gas in the draught tube region and medium-calorific value gas in the annulus region can be obtained. To characterize the ICFBG, the steady state mathematical model is proposed based on the bed hydrodynamics, the reaction kinetics and pyrolysis data of the coal. The proposed model can predict the product gas composition, carbon conversion, cold gas efficiency, gas yield and calorific value with reasonable accuracy. Acknowledgements The authors acknowledge a grant-in-aid for research from the Ministry of Trade, Commerce and Energy, Korea. References [1] Gutierrez LA, Watkinson AP. Fuel 1982;61:133. [2] Watkinson AP, Cheng G, Prakash CB. Can J Chem Engng 1983;61:468. [3] Saffer M, Ocampo A, Laguerie C. Int Chem Engng 1988;28:46. [4] Chatterjee PK, Datta AB, Kundu KM. Can J Chem Engng 1995;73:204. [5] Lee JM, Kim YJ, Lee WJ, Kim SD. Energy—The Int J 1998;23:475. [6] Neogi D, Chang CC, Walawender WP, Fan LT. AIChE J 1986;32:17. [7] Lee WJ. PhD Dissertation, Korea Advanced Institute of Science and Technology, Korea, 1995. [8] Herguido J, Corella J, Gonzalez-Saiz J. Ind Engng Chem Res 1992;31:1274. [9] Corella J, Anzar MP, Delgado J, Aldea E. Ind Engng Chem Res 1991;30:2252. [10] Judd MR. In: Proceedings of the 2nd International Coal and Gas Conversion Conference, Pretoria, 1987:23. [11] Jeon SK, Lee WJ, Kim SD. In: Large JF, Laguerie C, editors. Proceedings of the 8th Engineering Foundation Conference on Fluidization, Tours, France, 1995:445.

77

[12] Kim YJ, Lee JM, Kim SD. Fuel 1997;76:1067. [13] Lee JM, Kim YJ, Kim SD. Appl Thermal Engng 1998;18:1013. [14] Berggren JC, Bjerle I, Eklund H, Karlsson H, Svensson O. Chem Engng Sci 1980;35:446. [15] Riley RK, Judd MR. Chem Engng Commun 1987;62:151. [16] Judd MR, Rudolph V. In: Ostergaard K, Sorensen A, editors. Proceedings of the 5th International Conference on Fluidization, Elsinore, Denmark, 1986:505. [17] Kim YT, Song BH, Kim SD. Chem Engng J 1997;66:105. [18] Ahn HS. MS Thesis, Korea Advanced Institute of Science and Technology, Korea, 1995. [19] Grace JR, Lim KS. In: Grace JR, Avidan AA, Knowlton TM, editors. Circulating fluidized beds, London: Chapman and Hall, 1997. p. 504. [20] Hastaoglu MA, Berruti F, Hassam MS. In: Basu P, Large JF, editors. Circulating fluidized bed technology, Vol. II. Oxford: Pergamon Press, 1988. p. 281. [21] Marmo L, Manna L, Rovero G. In: Large JF, Laguerie C, editors. Proceedings of the 8th Engineering Foundation Conference on Fluidization, Tours, France, 1995:369. [22] Arena U, Chirone R, D’Amore M, Miccio M, Salatino P. Powder Technol 1995;82:301. [23] Zhang L, Li TD, Zheng QY, Lu CD. In: Anthony EJ, editor. Proceedings of the 11th International Fluidized Bed Combustion Conference, Montreal, Canada, 1991:1289. [24] Hyppanen T, Lee YY, Rainio A. In: Basu P, Horio M, Hasatani M, editors. Circulating fluidized bed technology, Vol. III, Nagoya, Japan, 1990:563. [25] Schoenfelder H, Kruse M, Werther J. AIChE J 1996;42:1875. [26] Balzer G, Simonin O. In: Preto FDS, editor. Proceedings of the 14th International Fluidized Bed Combustion Conference, Vancouver, Canada, 1997:1017. [27] Yoon H, Wei J, Denn MM. AIChE J 1978;24:885. [28] Weimer AW, Clough DE. Chem Engng Sci 1981;36:549. [29] Ma RP, Felder RM, Ferrell JK. Fuel Process Technol 1988;19:265. [30] Gururajan VS, Agarwal PK, Agnew JB. Chem Engng Res Design 1992;70:211. [31] Adanez J, Labiano FG. Ind Engng Chem Res 1990;29:2079. [32] Matsui I, Kunii D, Furusawa T. J Chem Engng Japan 1985;18:105. [33] Arthur JR. Trans Faraday Soc 1951;47:164. [34] Bhagat PM. Combust Flame 1980;37:275. [35] Haslam RT. Ind Engng Chem 1923;15:679. [36] Tesner PA. In: Proceedings of the Eighth Symposium on Combustion, 1960:807. [37] Batchelder HR, Busche RM, Armstrong WP. Ind Engng Chem 1953;45:1856. [38] Loison R, Chauvin R. Chemie et Industrie 1964;91:269. [39] Caram HS, Fuentes C. Ind Engng Chem Fundam 1982;21:464. [40] Chowdhury R, Chakravarty M, Bhattacharya P. Int J Energy Res 1991;15:593. [41] Hobbs ML, Radulovic PT, Smoot LD. AIChE J 1992;38:681. [42] Wen CY, Yu YH. AIChE J 1966;12:610. [43] Hovmand S, Davidson JF. In: Davidson JF, Harrison D, editors. Fluidization, New York: Academic Press, 1971. p. 193. [44] Kobayashi H, Arai H. Kagaku Kogaku 1967;31:239. [45] Foong SK, Lim CJ, Watkinson AP. Can J Chem Engng 1980;58:84. [46] Kikuchi K, Suzuki A, Mochizuki T, Endo S, Imai E, Tanji Y. Fuel 1985;64:368. [47] Lee JM, Kim YJ, Lee WJ, Kim SD. Hwahak Konghak 1997;35:121. [48] Foong SK, Cheng G, Watkinson AP. Can J Chem Engng 1981;59:625.