Coal gasification in a pilot scale fluidized bed reactor. 2. Gasification of

Apr 1, 1984 - Coal gasification in a pilot scale fluidized bed reactor. 2. Gasification of a New Mexico subbituminous coal. Mark J. Purdy, Richard M. ...
0 downloads 0 Views 742KB Size
Ind. Eng. Chem. Process Des. Dev. 7904, 23, 207-294

287

Coal Gasification in a Pilot Scale Fluidized Bed Reactor. 2. Gasification of a New Mexico Subbituminous Coal Mark J. Purdy, Richard M. Felder,' and James K. Ferrell Department of Chemical Engineering, North Carollna State University, Raleigh, North Carollna 27650

A New Mexico subbituminous coal was gasified with steam and oxygen in a 15.2 cm i.d. fluidized bed reactor at a pressure of 790 kPa (100 psig) and average bed temperatures between 875 and 990 OC. Material balances were obtained on total mass and major elements (C, H, 0, N, S). A simple representation of coal pyrolysis has been added to a previously developed model of gasification and combustion; the resulting model provides good correlations of measured carbon conversions, make gas production rates, and make gas compositions. Approximations that c a n be used to estimate sulfur conversion and the split between H,S and COS in the product gas have also been developed.

Introduction As a part of a continuing research program on the environmental aspects of fuel conversion, the U.S.Environmental Protection Agency has sponsored a research project on coal gasification a t North Carolina State University. The overall objective of the project is to characterize the gaseous and condensed phase emissions from the gasification-gas cleaning process and to determine how emission rates of various pollutants depend on adjustable process parameters. The facility used for this research is a small coal gasification-gas cleaning pilot plant which includes a continuous fluidized bed reactor, a cyclone separator, a venturi scrubber, absorption and stripping towers, and a flash tank for acid gas removal. Process control and data acquisition and logging systems and an extensive analytical laboratory complete the facility. A description of the pilot plant and operating procedures is given by Ferrell et al. (1980) and in abbreviated form by Felder et al. (1980). In experiments conducted to date, a devolatilized Western Kentucky bituminous coal and a raw New Mexico subbituminous coal have been gasified with steam and oxygen. In a previous paper (Purdy et al., 1981), a flow chart of the system is presented, the experimental procedures are outlined, and results of the runs with the char are summarized. (See also Ferrell et al., 1982a). The present paper summarizes the results of the subbituminous coal gasification studies. Additional details are given in a report by Ferrell et al. (1982b). Experimental Section Operating Procedures and Conditions. The following paragraphs summarize the procedures used in routine operation of the gasifier and the particular conditions of the runs for which data will be presented in this paper. Additional details about the system may be found in previous publications (Felder et al., 1980; Ferrell et al., 1980; Purdy et al., 1981), and more detailed descriptions of the analytical devices and procedures used for gaseous and condensed phase effluents are given by Ferrell et al. (198213). The pulverized coal feed stream is metered into the top of the reactor (a 15.2 cm i.d. schedule 40 stainless steel pipe) from a pressurized hopper by controlling the rate of revolution of a feed screw conveyor. Feed gases enter the bottom of the reactor through three triangularly spaced feed nozzles. Oxygen is fed from a bank of six cylinders, 0196-4305/04/1123-0207$01.50/0

nitrogen from a 6000-gal liquid nitrogen tank with an atmospheric vaporizer, and steam from a process steam line. All gas feed steams are preheated. Spent char is removed from the bottom of the bed through a screw conveyor into a second pressurized hopper. Temperatures in the bed are monitored with six thermocouples located at various vertical positions within a central tube in the reactor. Taps at several points in the reactor are used to obtain absolute and differential pressure measurements. The level of the fluidized bed is monitored with a nuclear level gauge and is controlled by adjusting the spent char removal screw rotation rate. The make gases leaving the reactor pass through a cyclone separator and then through a venturi scrubber, which together knock out mwt particulates, tars and condensable species, and water-soluble species. The condensate from the scrubber is collected in a receiving tank. Multiple spent char samples are taken from the char receiver vessel and from the bottom of the reactor after a run, and samples of elutriated dust are taken from the cyclone collection tank. Analyses performed on these samples and on the feed coal include proximate and ultimate analysis and analysis for selected trace elements. Gas samples for chromatographic analysis are taken at operating pressure at a point just downstream of the cyclone in l-L stainless steel bombs, and at lower pressures with l-L glass bombs. The glass bombs are coated internally with hexamethyldisilizaneto reduce the adsorption of trace sulfur, nitrogen, and hydrocarbon species. A small side stream is drawn off just downstream of the cyclone and passes through a sample train. During steady-state operation, the gas flow through the sample train is metered with a dry test meter, and the water condensed in a water-cooled trap is weighed. The result provides the primary measurement of the water content of the make gas. A packed cold tar trap provides samples of particulates, tars, and condensable species in the make gas. In the runs performed in this study, the coal fed to the gasifier was a New Mexico subbituminous coal (Navaho Mine, Fruitland, NM), ground to a 10 X 80 mesh size (average diameter = 700 pm). Multiple samples of the feed coal were taken before each run, sieved, and subjected to proximate and ultimate analyses, and analyses for selected trace constituents. Representative proximate and ultimate analyses of the coal are given in Table I. Gasification runs were carried out at 14 sets of process 0 1984 American Chemical Society

288

Ind. Eng. Chem. Process Des. Dev., Vol. 23, No. 2, 1984

Table I. Analysis-Navaho Mine Subbituminous Coal, Fruitland, NM (Weight Percent)

___

Proximate Analysis fixed carbon volatile matter moisture ash

36.2 31.1 9.7 23.0 m

Ultimate Analysis carbon hydrogen oxygen nitrogen sulfur ash

D D

50.2 4.2 20.7 1.1 0.8 23.0

conditions. (See Table 11). Run conditions were established by choosing the desired average bed temperature, coal feed rate, and molar steam-to-carbon feed ratio. The average fluidized bed temperature ranged from 875 to 990 OC, and the molar steam-to-carbon ratio ranged from 0.97 to 1.90. The coal feed rate was varied from 18 to 30 kg/h (40-67 lb/h). Oxygen was fed as needed to maintain the desired bed temperature, and nitrogen was used to maintain the inlet gas velocity at approximately 0.3 m/s (1.0 ft/s). The pressure was held at roughly 790 kPa (115 psia) and the feed gas temperature was held constant at 540 OC (lo00 OF). All runs were at a bed height of 0.97 m (38 in.). Balances. Material balances were obtained on total mass and on major elements (C, H, 0, N, S). Table I1 shows the percentage recoveries of total mass and the worst of the percentage recoveries of the elements C, H, and 0. Mass recoveries were generally between 95% and 105%, validating both the experimental procedures used to obtain the data and the conclusions derived from the data in the modeling studies. Sulfur recoveries, which were based on measurements of very low concentrations of gaseous and condensed phase species, varied to a much greater extent, but with only one exception were between 80% and 120%. Solid and Tar Effluents. The elutriation rate from the reactor was determined from measurements of the particulates in the cyclone separator and sample train cold trap. CalcuIated elutriation rates ranged from 1.0 to 2.2 kg/h (2.2-4.9 lb/h), or 5 to 11% of the coal feed rate. The elutriates had a high ash (30-70%) and carbon (2045%) content, but only small amounts of hydrogen, oxygen, and nitrogen. The carbon content of the elutriated material represented 3 to 9% of the feed carbon. The char removal rate from the gasifier ranged from 2 to 14 kg/h (4-30 lb/h). Compared to the elutriates, the char composition was generally slightly higher in ash ( 4 0 4 5 % ) and lower in carbon content (lo-%%). The sample train cold trap also provided data from which the tar production rate was calculated. Tar rates ranged from 1.5 to 2.2 kg/h (3.3-4.9 lb/h), or 10-15% of the dry ash-free coal feed rate. Roughly 10 to 16% of the feed carbon was evolved in the tar. Effects of Operating Parameters. The effect of the average bed temperature on the dry N2-freemake gas flow rate is shown in Figure 1. The average bed temperature used here is a volume-weighted average, calculated from thermocouple measurements at four locations within the fluidized bed. Figure 2 shows the effect of the average bed temperature on carbon conversion. Both carbon conversion and make-gas production increase with increasing temperature, as would be expected from the known positive effect of temperature on the degree of devolatilization and rates of the gasification reactions. The effects of steam-to-carbon feed ratio cannot be seen directly from the data, since the temperature could not be

. 0

D

1550 1650 1750 Average Bed T e m p e r a t u r e , F

IS50

Figure 1. Effect of temperature on dry make gas rate.

Ind. Eng. Chem. Process Des. Dev., Vol. 23,No. 2, 1984 289

Table 11. Summary of Coal Gasification Run Data

___

_ _ _ _ i _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

a

run

T. "C

P. kPa

66 68 B 69 B 72 74 76 78 79 80B 81B 81D 82 84 85B

981.4 974.7 9 37.9 980.7 903.8 933.1 874.4 953.1 962.5 936.3 889.2 987.5 924.9 974.6

815.9 825.2 830.8 821.4 828.4 794.9 817.5 815.0 814.3 818.2 814.2 827.3 829.8 824.0

run

% C conv

dry n;ake gas rate, SCMM

66 68 B 69 B 72 74 76 78 79 80 B 81B 81D 82 84 85B

90.8 68.7 71.4 84.6 42.2 79.0 64.2 79.2 78.9 72.7 65.9 89.8 71.8 84.2

0.56 0 -66 0.53 0.55 0.45 0.54 0.45 0.53 0.49 0.55 0.47 0.55 0.46 0.61

Feed ratio.

__

mol of H,O/ mol of ca 1.53 1.12 1.56 1.72 0.97 1.54 1.71 1.73 1.51 1.21 1.69 1.90 1.82 1.62

sweet gas H.V., kJ/SCM

1 5 371 1 4 079 1 3 847 13959 1 3 988 1 4 043 1 4 679 1 3 567 1 39 4 3 1 4 238 1 4 959 1 4 314 1 4 499 1 4 433

mol of O,/ mol of Ca

__

0.30 0.20 0.24 0.23 0.10 0.19 0.15 0.20 0.20 0.17 0.16 0.22 0.19 0.19 5% mass re cover y

95.5 102.5 97.7 101.2 98.0 98.6 102.0 102.4 92.2 93.7 98.8 96.1 100.0 96.8

solids res. time, min

___-____17.3 9.9 9.2 21.5 9.4 12.3 16.0 12.1 9 .o 13.7 13.9 9.5 8.2 18.5 worst

% recovery

107.3(C) 108.5(C) 106.8(H) 110 .O(C) 96.6(H) 109.2(C) 109.1(C) 105.5(0) 107.9 (H) 92.6(0) 104.1(H) 105.4(C) 101.9(H) 106.8 (0)

Heating value of sweet gas ( H , , CO, and CH, in make gas).

The "Johnson kinetics" reaction sequence is used to describe the steam/carbon gasification process: reaction of steam with carbon to form carbon monoxide and hydrogen, hydrogenation of carbon to form methane, and a third reaction which is the stoichiometric sum of the first two. The "char reactivity coefficient", a parameter that accounts for differences in intrinsic reactivities of different chars, is the only adjustable model parameter associated with the gasification process; rate and equilibrium constants for the three process reactions are those given by Johnson (1974). The final reaction in the combustion/ gasification sequence is the water gas shift. A rate expression given by Wen and Tseng (1979) is used, with the rate constant being the third adjustable parameter of the model. Purdy et al. (1981) used this model to correlate data on the gasification of a devolatilized Western Kentucky bituminous coal. They showed that despite its simplicity, the model does a good job of predicting carbon conversion, make gas production rate, and make gas composition for varying average bed temperature and steam-to-carbon feed ratio. Moreover, they show that the incorporation of water gas shift kinetics into the model provides significantly better correlations than are obtained if the common assumption of shift reaction equilibrium is made. For the model to be applicable to the gasification of a raw coal, the pyrolysis or devolatilization stage of the gasification process must be incorporated. It is assumed that devolatilization reaches equilibrium between the reactor inlet and the fluidized bed. This assumption is justified by numerous studies which show that at representative fluidized bed gasification temperatures (900-10o0 "C),the pyrolysis process is complete in times on the order of tens of milliseconds (Anthony and Howard, 1976; Agreda et al., 1979). An expression derived from results obtained in bench-scale pyrolysis experiments performed in our labo-

ratory (Kau, 1980) is used for the fractional weight loss at equilibrium on a dry ash-free basis fd

= 1.052 - 0.7150(2'/100) + 0.1615(T/100)2 -1.367 X 10-2(~/100)3+ 4.044 x 10-4~~/100)4~(1)

where T is the temperature in Kelvins. In the current form of the model, the composition of the devolatilization product gas is taken from the data of Friedman et al. (1968), who conducted pyrolysis research on a Wyoming subbituminous coal similar to that used in this research. No sulfur gas analyses were performed in the study; therefore the results of separately conducted pyrolysis experiments (Gilman, unpublished results) have been used to obtain data on the species H2S and COS. Based on the two studies, the gas evolved from the New Mexico subbituminous coal is taken to contain 19% carbon monoxide, 28% hydrogen, 37% methane, 16% carbon dioxide, 0.2% hydrogen sulfide, and 0.014% carbonyl sulfide, and is assumed to be independent of process conditions. Tars are estimated by using the tar to total volatiles ratio of 0.34 measured by Friedman. It is assumed that water is evolved only by the devolatilization of the moisture fraction of the coal as determined by proximate analysis. When the model is implemented, the first step is to simulate the devolatilization of the feed coal, using the formulas given above. The residual char from this step is taken to be the feed to the fluidized bed, where the gasification and combustion steps take place. The simulation of these two steps and the estimation of the three adjustable model parameters proceed as previously outlined (Purdy et al., 1981). Parameter Estimation. The model as presently constituted has three adjustable parameters: the coal reactivity coefficient,fo, the combustion product distribution

290

Ind. Eng. Chem. Process Des. Dev., Vol. 23, No. 2, 1984

0

20

40

BO

50

100

0

2

Figure 3. Predicted vs. experimental carbon conversion.

3 r y Gas R a t e ( s t f m l

Figure 5. Predicted vs. experimental H2production rate.

/

15

20

25

Expertmental

20

10

0 10

5

3

Experimental

Experimental

30

Expertmental

Figure 6. Predicted vs. experimental CO production rate.

Figure 4. Predicted vs. experimental dry make gas rate.

coefficient, a, and the water gas shift reactivity parameter, fw The runs with the best maas balance closures (GO-66, 68B, 69B, 71B, 72, and 73B) were chosen to provide the data base for the parameter estimation. The parameters were estimated by minimizing a weighted s u m of squares of normalized residuals of the make gas flow rate, the carbon conversion, and the product gas mole fractions of hydrogen, carbon monoxide, and carbon dioxide, using a Pattern Search algorithm to achieve the minimization. Details of the procedure are given by Purdy et al. (1981). The parameters estimated in this manner are fo = 4.20, a = 1.00, and fwg = 1.20 X lo4. The estimated value of a indicates that carbon oxidized forms only C02. (This should not be taken to be a mechanistic statement: the combustion product distribution may be influenced by secondary reactions so that the resulting distribution is not indicative of the initial products.) In several gasification studies a value of a = 1.0 has been assumed (Caram et al., 1979; Institute of Gas Technology, 1978; Pukanic et al., 1978; Sundaresan and Amundson, 1979), an assumption supported by the result just given. The estimated value off, indicates that the shift reaction rate is approximately four orders of magnitude less than the rate typically obtained in catalytic shift reactors. Wen and Tseng (1979) used a value of 1.7 X lo4 in modeling the gasification of a bituminous coal by the SYNTHANE process. The greater value used by Wen and Tseng may easily be attributed to the differences between the coals used in the two studies.

C02 Rate ; l b i n r )

M

020 N

25

30

35

40 0

Experlmental

Figure 7. Predicted vs. experimental CO, production rate.

Data Correlations. The model was used to simulate all gasifier runs shown in Table 11. Plots of predicted vs. measured values of carbon conversion and dry make gas flow rate are shown in Figures 3 and 4. The reasonably close proximity of the points to the 45' line is gratifying in view of the simplicity of the model. The model also does a creditable job of correlating data on the evolution of individual species. Figures 5-8 show predicted vs. measured rates of evolution of H2, CO, and CHI from the gasifier. The correspondence between prediction and measurement suggests that the mode1 can be used to predict the composition of the gasifier make gas

Ind. Eng. Chem. Process Des. Dev., Vol. 23, No. 2, 1984

Experlmental

291

Experlmental

Figure 8. Predicted vs. experimental CHI production rate.

Figure 10. Equilibrium va. experimental K value.

'I sa

/ 0

0.3

0.4

0.5

0.6

0.7

I

Experimental

Figure 9. Predicted vs. experimental K value. 4a

-

for a specified set of reactor conditions and also to study the effects of individual reactor variables on product yield. Figure 9 shows predicted vs. measured value of the ratio This ratio would equal the water gas shift equilibrium constant if the reaction actually proceeded to equilibrium in the fluidized bed. A high degree of scatter is observed in the plot, reflecting the fact that errors in any of the composition variables in K lead to additional errors in the other variable values, thereby augmenting the error in K. In Figure 10, the values of K corresponding to equilibrium at the gasifier exit temperature are plotted against the experimental values of this ratio. The clear tendency of the points to fall above the 4 5 O line, especially relative to the corresponding points on Figure 9, shows the improved correlation that can be obtained by taking water gas shift reaction kinetics into account rather than assuming equilibrium. The make gas heating value was calculated from experimental results and from model predictions for each run. The detailed results are given by Ferreu et al. (1982b). The error between predicted and measured heating values was less than 5% for all runs with good mass balance closures. Effects of Operating Variables on Reactor Performance. The model was run for several hypothetical reactor conditions to determine the effects on reactor performance of selected operating variables. Figures 11 and 12 show the effects of the average bed temperature

I

1500

1000 T.mp.r.ture.

1100

1800

J

OF

Figure 12. Predicted effect of gasifier temperature on dry make gas rate.

on carbon conversion and make gas production rate for fiied steam-to-carbon feed ratios; Figures 13 and 14 show

292

Ind. Eng. Chem. Process Des. Dev., Val. 23, No. 2, 1984

E c'

l

D

L

a C >

0 0 C

P 0

76 1.o

1.5

2.5

20

30

Molar Steam-lo-Carbon R a t l o

Figure 13. Predicted effect of steam-to-carbon feed ratio on carbon conversion.

100

300

500

700

Pressure. PEIB

Figure 15. Predicted effect of pressure on carbon conversion.

221

T = 16OO0F

I

1.0

15

20

25

30

Molar Steam-lo-Carbon Rallo

Figure 14. Predicted effect of steam-to-carbon feed ratio on dry make gas rate.

the effect of the steam-to-carbon ratio at fixed temperatures; Figures 15-18 show the effect of pressure at fixed temperatures; and Figures 19 and 20 show the effect of solid-phase space time (bed holdup divided by solid feed rate) with bed temperature and coal and steam feed rates held constant (i.e., varying the bed depth). Figures 11 and 12 show, as would be expected, that carbon conversion and make gas flow rate both increase sharply with increasing temperature (see Figure 1). The plots also indicate that at low temperatures the steamto-carbon ratio has little effect on either performance variable. The latter conclusion is corroborated by Figures 13 and 14, which also show that at higher temperatures increasing the steam-to-carbon ratio increases both carbon conversion and make gas production, undoubtedly due in part to the dependence of the rate of the steam/carbon reaction on the steam partial pressure. Figures 15-18 show that below 200 psia increasing pressure increases carbon conversion, make gas flow rate, methane production, and the shift reaction K value. Above 200 psia, much smaller effects are seen on carbon conversion and make gas flow, while methane production continues to increase with increasing pressure. Shift

I

L

100

500

300

700

Pressure, p ~ l a

Figure 16. Predicted effect of prssure on dry make gas rate.

100

300

500

700

Pressure. psla

Figure 17. Predicted effect of pressure on methane production rate.

Ind. Eng. Chem. Process Des. Dev., Vol. 23,No. 2, 1984

293

L-7 T=1600°F

500

300

100

700

Prerrure, pain

Figure 18. Predicted effect of pressure on K value.

0

5

10

25

20

15

30

Solid-Phase Space T i m e , m ~ n

Figure 20. Predicted effect of solid-phase space time on carbon conversion.

X C Conuers 1 on 0

5

10

15

20

25

30

Solid-Phase Space T i m e , mln

Figure 21. Sulfur conversion vs. carbon conversion.

Figure 19. Predicted effect of solid-phase space time on carbon conversion.

equilibrium is apparently attained at pressures over 200 psia. Figures 19 and 20 indicate an expected positive dependence of carbon conversion and make gas production rate on solid-phase space time. The rates of production of individual species show similar monotonic dependences on space time.

Sulfur Gas Evolution Previous studies with a char feed (Ferrell et al., 1982a) indicated that the sulfur conversion in the fluidized bed could be roughly estimated by assuming it equal to the carbon conversion. Figure 21 plots sulfur conversion vs. carbon conversion for the runs in the present study. It appears from this plot that the equation of the two conversions is also a reasonable first approximation for gasification of the New Mexico subbituminous coal. The principal produds of sulfur conversion are hydrogen sulfide and carbonyl sulfide. Data obtained in the previous study of char gasification (Ferrell et al., 1982a) suggested that the reaction COS + HzO = HzS + COz (2) could be taken to proceed to equilibrium at the reactor

1400

1500

1600

1800

1700

T e m p e r a t u r e at l o p o f Bed,

300

F

Figure 22. Comparison of experimental values of K. with data of Kohl and Riesenfeld.

outlet temperature. Figure 22 plots experimental values of the logarithm of the ratio (3) K , = [CO,l [HZSI /[H,OI [COS1 vs. inverse absolute temperature. Also shown on the plot is a correlation for the equilibrium constant based on data of Kohl and Riesenfeld (1979) In K, = -1.352 + (7881.6)/T(OR) (4) The scatter is considerable, again reflecting the fact that

294

Ind. Eng. Chem. Process Des. Dev., Vol. 23,No. 2, 1984

errors in measurement of any concentration lead to errors in other calculated concentrations, and hence magnify the error in K,. However, the fact that the points scatter about the correlation line supports the idea that the given reaction may be considered at equilibrium to obtain a rough estimate of the H2S/COS split. The results given above have led to the incorporation into the model of an algorithm for predicting H2S and COS emissions for a given set of reactor operating conditions. The carbon conversion is first estimated as described previously. The total sulfur conversion is next calculated by equating it to the carbon conversion. Finally, the H2S/COS shift reaction equilibrium constant is calculated from eq 4 and is used to determine the split between these two species. From the known total sulfur gas emission and the known division between H2S and COS,the emission rates of the two sulfur gases are calculated. Work is currently underway to model the kinetics of sulfur gas evolution. Improved predictions of the H2S and COS emission rates should be obtained when the results of these studies are incorporated into the gasifier model. Summary and Conclusions A New Mexico subbituminous coal was gasified with steam and oxygen in a fluidized bed reactor at varying temperatures and steam-to-carbon feed ratios. Carbon conversion and make gas production rates and compositions were correlated with a three-stage reactor model. As was the case in the previous study of the gasification of a devolatilized bituminous coal, the model, despite its relative simplicity, does a good job of correlating the process data. The incorporation of water gas shift reaction kinetics into the model gives improved results compared to the assumption of shift equilibrium. Although the three adjustable model parameters must be determined separately for each new coal feedstock, the model may thereafter be used to correlate data for a wide variety of coals and to serve as a basis for preliminary design and optimization. The average bed temperature is the single most important variable affecting the make gas production rate. Carbon conversion and the make gas production rate both increase sharply with increasing temperature. At the lower temperatures investigated the steam-to-carbon feed ratio has a relatively slight effect on either performance variable, while at higher temperatures both variables increase with increasing steam-to-carbon ratio. The model predicts that increasing the pressure increases carbon conversion, the make gas flow rate, and methane production. The effect is most dramatic for prssures under 200 psia. Above 200 psia the effect on carbon conversion and the make gas rate is small, while methane production continues to increase, and shift reaction equilibrium is approached. The sulfur conversion in the gasifier is approximately equal to the carbon conversion. The distribution of the emitted sulfur gases between hydrogen sulfide and carbonyl sulfide may be estimated by assuming that the COS-H20/H2S-C02 reaction reaches equilibrium at the reactor outlet temperature.

Future papers in this series will report on the steam/ oxygen gasification of a North Carolina peat and on the application of more sophisticated kinetic and hydrodynamic models to the conversion data for all feedstocks studied. Acknowledgment This research was supported in part by Environmental Protection Agency Cooperative Agreement No. CR804811 and in part by Department of Energy Grant DE-FG2280PC-30-232-AO02. The authors acknowledge with gratitude the contributions to the work of Professor Ronald Rousseau, Bill Willis, Gary Folsom, Kathy Steinsberger, Paula Jay, Netter Murphy, and Terrie Staley.

Nomenclature a = combustion product distribution coefficient f d = devolatilization fractional weight loss on a dry ash-free basis at equilibrium fo = coal reactivity coefficient f,, = shift reactivity parameter K = equilibrium constant for the water gas shift reaction K, = equilibrium constant for the H2S/COS shift reaction P = reactor pressure T = average temperature in fluidized bed Tout= temperature at outlet of fluidized bed Registry NO.02,778244-7; H2,1333-74-0;CO, 630-08-0;C02, 124-38-9; CHI, 74-82-8; H2S, 7783-06-4; COS, 463-58-1.

Literature Cited Agreda, V. H.;Feider, R. M.; Ferreil, J. K. "Devolatlllzatlon Kinetics and Elemental Release in the Pyrolysis of Pulverized Coal"; EPA-600/7-79-241, 1979. Anthony, D. 0.; Howard, J. 0 . AlChE J . 1976, 22, 625. Caram, H. S.; Amundson, N. R. I n d . Eng. Chem. Process Des. Dev. 1979, 13, 80. Felder. R. M.; Kelly, R. M.; Ferrell, J. K.; Rousseau, R. W. Environ. Sei. Techno/. 1980, 14, 858. Ferrell, J. K.; Felder, R. M.; Rousseau, R. W.; McCue, J. C.; Kelly, R. M.; Wl111s, W. E. "Coal GaslflcationlGas Cleanup Test Facility: Vol. I. Description and Operation"; EPA-800/7-80-046a. 1980. Ferrell. J. K.; Felder, R. M.; Rousseau. R. W.; Ganesan, S.; Kelly, R. M.; McCue, J. C.; Purdy, M. J. "Coal Gaslflcatlon/Gas Cleanup Test Facility: Vol. 11. Environmental Assessment of Operation with Devolatlllzed Bituminous Coal and Chilled Methanol"; EPA-600/7-82-023, 1982a. Ferrell, J. K.; Feider, R. M.; Rousseau, R. W.; Kelly, R. M.; Purdy, M. J.; Ganesan, s. "Coal GaslficationlGas Cleanup Test Facility: Vol. 111. Envlronmental Assessment of Operation with New Mexico Subbituminous Coal and Chilled Methanol"; EPA-600/S7-82-054, 1982b. Friedman, L. D.; Rau, E.; Eddinger, R. T. Fuel 1968, 47, 149. Institute of Gas Technology "Coal Conversion Systems Technical Data Book"; prepared for US. Department of Energy, Contract No. EX-76-C01-2286, Report No. HCP/T228&01,UC90, 1978. Johnson, J. L. A&. Chem. Ser. 1974, No. 131. Kau. C. C., M.S. Thesis, N.C. State University, Raleigh, NC, 1980. Kohl, A. L.; Rlesenfeld, F. C. "Gas Purification", 3rd ed.; Gulf: Houston, TX, 1979. Pukanic, 0. W.; Cobb, J. T.; McMlchael, W. J.; Haynes, W. P.; Strakey, J. P. "Mathematical Modeling of the SYNTHANE Gasifier"; presented at 7151 Annual AIChE Meetlng, Miami, FL, 1978. Purdy, M. J.; Felder, R. M.; Ferrell, J. K. I d . Eng Chem. Process D e s . Dev. 1981, 20, 675. Sundaresan, S.;Amundson, N. R. Chem. Eng. Sci. 1979, 3 4 , 345. Wen, C. P.; Tseng, H. P. "A Model for Fluidized Bed Coal Gasification Simulation"; presented at 72nd Annual AIChE Meeting, San Francisco, CA, 1979.

.

Received for review April 15, 1982 Revised manuscript received April 1, 1983 Accepted August 22, 1983