Ind. Eng. Chem. Process Des. Dev., Vol. 17, No. 2, 1978 Guin, J. A., Tarrer, A. R., Taylor, L., Prather, J., Green, S., Ind. Eng. Chem. Process Des. Dev., 15 (4),490 (1976). Henley, J. P., Guin, J. A., Tarrer, A. R., Pitts, W. S., Styles, G. A,, Prather. J. A,, Am. Chem. SOC.Div. Fuel Chem. Prepr., 21 ( 5 ) 59 (1976). Hikita, H., Kikukawa. H., J. Chem. Eng. Jpn., 8 (5),412-413 (1975). Hikita, H., Kikukawa. H., Chem. Eng. J., 8, 191-197 (1974). Hughmark, G. A.. Ind. Eng. Chem. Process Des. Dev. 6, 218-220 (1967). Kato. Y., et al., Int. Chem. Eng., 13 (3),562 (1973). Mashelkar, R. A., Sharma, M. M., Trans., Inst. Chem. Eng., 48, TI62 (1970). Mixon, F. O.,Whitaker, D. R., Orcutt, J. C., AlChEJ., 13, 21 (1967). Ostergaard. K., Suchozebrski, W., Proceedings of the Fourth European Symposium, Chemical Reaction Engineering, 1968. Pavlica. R. T., Olson, J. H., lnd. Eng. Chem., 62 (12),45-58 (1970). Pitts, W. S.,Ill, M. S. Thesis, Auburn University, 1976. Prather. J. W., Tarrer, A. R., Johnson, D. R., Guin, J. A,, Am. Chem. SOC.Div. Fuel Chem. Prepr., 21 (9,144 (1976).
135
Reith, T., Renken, S.,Israel, B. A., Chem. Eng. Sci., 23, 619 (1968). SRC Technical Report No. 8, Catalytic Inc. Philadelphia, Pa., 1975. Voyer. R. D.. Miller, A. I., Can. J. Chem. Eng., 46, 335 (1968). Wen, C. Y., Han, K. W., Am. Chem. SOC.Div. FuelChem. Prepr., 20, (1)216
(1975). Wehner, J. F., Wilhelm, R. H., Chem. Eng. Sci., 6, 89-93 (1956). Yoshida, F., Akita, K., AIChEJ., 11, 9 (1965).
Received for review April 8, 1977 Accepted October 11, 1977
This work was supported by the U. S. Energy Research and Development Administration under Contract No. Ex-76-S-01-2454.
Particle Separation from a Fluidized Mixture. Simulation of the Westinghouse Coal Gasification Combustor/Gasifier Operation Joseph L.-P. Chen and Dale L. Keairns* WestinghouseResearch Laboratories,Pittsburgh, Pennsylvania 15235
Particle separation data from a 101.6-mm pressurized (up to 660 kPa) model operating at low gas velocity, near the Umfof the mixture, and from a 114-mm Plexiglas model operating at high gas velocity, near the Utof the mixture, with a nozzle located at the center of a conical bottom, are presented. These models were used to simulate the design and operation of alternative fluidized-bed agglomerating combustor/gasifier concepts. Comparison of the data obtained from these two modes of operation is made and a design for the separation of the agglomerated ash at low gas velocity is established. The data indicate that >95% separation of the agglomerated ash from char at a rate of about 400 kg/min m2 of the separator cross-sectional area can be achieved if the operating gas velocity is close to the minimum fluidizing velocity, e.g., 1.0 to 1.5 times Umc,of the agglomerated ash.
Introduction Particle separation data were used to provide design and operating criteria for a multistage fluidized-bed coal gasification process being developed at Westinghouse (Archer et al., 1973) to produce low-Btu gas for combined-cycle power generation. An experimental investigation of the separation of agglomerated ash in a batch fluidized char-ash mixture at low gas velocity has been previously reported (Chen and Keairns, 1975). Good particle separation was obtained for an operating gas velocity close to the minimum fluidizing velocity of the separated species. In order to simulate directly the char-ash separation in the combustor/gasifier and to investigate alternative processes for separating agglomerated ash from the mixture, experiments were carried out to investigate continuous particle separation. Two processes were investigated. One was to separate the agglomerated ash from the fluidized bed by using a low gas velocity near the minimum fluidizing velocity of the ash, and the other was to use a high gas velocity near the terminal velocity of the char. Results from these two experiments are compared. The data are used to project operating conditions for achieving the desired particle separation in the fluidized-bed coal gasification combustor/gasifier. A 15-tonper-day process development unit (PDU) is now being operated which incorporates char-ash separation based on the use of minimum fluidizing velocity. Literature Segregation of two different solid particles with appreciable density difference in a low velocity fluidized bed has been reported by the British Gas Council in their recirculating coal 0019-7882/78/1117-Ol35$01.00/0
gasification process (Cockerham, 1962) and by the Consolidation Coal Co. in their COz acceptor gasification process (1968). While the British Gas Council has achieved about 80 to 90% purity of ash separated from their agglomerating gasifier, which was fluidized a t about four times the U,f of the ash mixture, Consolidation Coal Co. has achieved a much higher separation efficiency (>95%) for their spent stone (lime or dolomite) in the COZ acceptor process. Previous investigations a t Westinghouse (Chen and Keairns, 1975) indicated that a high separation efficiency, e.g., >95%, can be obtained for the agglomerated ash if a gas velocity close to the U,f of the agglomerated ash is used. The conclusion on the optimum gas velocity is consistent with work reported by Rowe et al. (1972b). Jequier et al. (1960) experimented with an alternative concept for coal gasification, ash agglomeration, and particle separation. The apparatus consisted of a fluidized chamber a t the top and a conical duct between the chamber and a venturi. A high gas velocity prevailed a t the venturi, which allowed particles of agglomerated ash to fall through but not the fine char. For a gas velocity a t the venturi in the order of 15 m/s, they showed that the separated mixture contained about 95% ash. Separation of agglomerated ash from the char-ash mixture was not smooth, however, and the efficiency decreased rapidly when the gas velocity at the venturi decreased. Their data also show that an increase of the gas velocity a t the venturi was required in order to keep coal fines from falling through the nozzle as the bed height was increased. Experimental Apparatus and Procedures Two fluidization cold models were used. One was for sepa-
0 1978 American Chemical Society
136
Ind. Eng. Chem. Process Des. Dev., Vol. 17, No. 2, 1978
Table I. Char-Dolomite Mixtures Used in the Combustor-Gasifier Cold Models to Simulate Char-Ash Separation Material
Size, pm
Density,a glc m3
Umf, mls
Ut,b mls
Mixture A Char -0.8-1.2 4.06-0.12 -1.74 841 X 420 Dolomited 1410 X 841 -2.88 -0.64 -6.9 Mixture B Char -0.8-1.2 4.07-0.14 -1.8 841 X 595 Dolomite -2.88 -0.6 -6.7 1190 X 841 Mixture C Char -0.8-1.2 -0.07-0.15 -1.8 55% 1410 X 841,45% 595 X 420 Dolomite 40% 2000 X 1410,60%841 X 595 -2.88 -0.43 -4.8 Calculated from bulk density. A t 101 t o 170 kPa and 25.6 O C . Projected char size for the PDU: -6000 pm. Dolomite used to simulate ash agglomerates. Uniform and coarse (-3000 to 6000 pm) agglomerates have been obtained by IGT (Lee and Tarman, 1974) in their agglomerating gasifier.
-r 0.61 m
t
,Glass
0.61 r
Glass
102 mm I O 0 61 m tong
Ro t a r y
Feeder
Receiving Hopper
Alr
Figure 1. Pressurized cold model simulating the combustor/gasifier.
ration of particles a t low gas velocity and the other for separation a t high gas velocity. Three mixtures which simulated the char-ash mixtures in the agglomerating combustorlgasifier were used to study the degree and rate of separation. A dolomite which was slightly heavier than the agglomerated ash (Lee and Tarman, 1974) was used to simulate the agglomerates. The mixtures consisted of two-component and wide-size distribution particles of different size and density. Table I shows the property of these materials. The actual particle sizes and densities present in the agglomerating combustor/gasifier will be determined in the PDU. The density difference of char and dolomite is expected to be comparable to that of char and agglomerated ash. Previous studies a t Westinghouse (Chen and Keairns, 1975) indicated that the separation process is largely controlled by the density difference of char and ash, especially when the gas velocity is close to or greater than the minimum fluidizing velocity of the largest ash particles a t high pressure. The results obtained from this study are therefore expected to be representative of the process development unit operation. Low Velocity Separation. Simulation of the separation of agglomerated ash from char a t low gas velocity, close to the
U,f of ash, was carried out in a 102-mm pressurized unit a t pressures up to 660 kPa. Figure 1shows the schematic diagram of the apparatus. The model was similar to that used for the batch experiments described in the previous paper (Chen and Keairns, 1975). The fluidized bed column was composed of two 102-mm diameter glass pipes and two 102-mm steel pipes. The unit was about 2.6 m high and was able to withstand pressure up to 1130 kPa. A 51-mm i.d. Plexiglas tube 0.762 m long was installed inside the lower glass section. A conical distributor (60’ cone angle with 48 0.8-mm holes arranged in three rings 6.35 mm apart) connected the Plexiglas tube to the glass pipe. The upper and lower columns simulated the gasifier and the agglomerating combustor, respectively. A 12.7-mm copper tube installed inside the combustor simulated the air inlet and formed the annulus for the char-ash separator. A distributor consisting of 32 0.8-mm holes arranged in two rings 6.35 mm apart was installed at the bottom of the annulus. All the distributors were covered with 80-pm screens to prevent fine particles from falling through or plugging the orifices. Independent air supplies were used to fluidize the particles above the conical distributor, in the center nozzle, and in the char-ash separator. A solid feeding system consisting of feed hoppers and a rotary feeder was used to feed the agglomerated ash continuously into the fluidized mixture a t a controlled rate. A solid withdrawal system, consisting of a rotary feeder and a receiving hopper, was used to remove the separated ash. The primary measurements included airflow rates, column pressure, and the concentration of char in the separated particles. The fluidized bed column was initially filled with the char-ash mixture and fluidized for about 5 min by controlled gas velocities in the simulated gasifier, agglomerating combustor, and the char-ash separator to induce a steadystate segregated particle distribution in the column. Dolomite was then continuously fed a t a controlled rate into the gasification section for about 3 min. At the same time, solids were withdrawn from the bed a t a rate calculated to maintain the char-rich and dolomite-rich interface near the top of the char-ash separator. Samples from the removed particles, excluding the quantity initially in the char-ash separator, were screened to determine their dolomite concentration as a function of the gas velocity and the system pressure. The rate of separation was determined by fluidizing pure char in the gasifier, the combustor, and the char-ash separator, and then feeding dolomite into the gasifier. The dolomite separated from the mixture and replaced char in the char-ash separator. The rate of char-ash separation was measured by determining the rising level of the char-ash interface in the char-ash separator. High-Velocity Separation. Separation of ash (simulated by dolomite) from a char-ash mixture by entrainment was
Ind. Eng. Chern. Process Des. Dev., Vol. 17, No. 2, 1978
Solid Loading Port-\
137
, 8 I
-@
To Manometer
I
I
0
0.61 m
0
50.8 mm
i-;
Umt of
93*0
4To Manometer 4To Manometer
92.0
Dolomite
,
0.5
!' v, 1 1
I
I
v2 o v 2 = 1 4 . 0 rn/sec, v 3 = 0.38 m / s e c v2 =8.8 m l s e c , v 3 = 0.115 m / s e fU, of C h a r 0.061 - 0.122 m / s e c Bed Material Char 841 x ;2l p m Bed M a t e r i a l D o l o m i t e 1410 x 841 prn in t h e C h a r / Pressure 134.7 k P a A s h S e p a r a t o r T e m p e r a t u r e 25.6'C "term slugs ' I of C h a r
\.I
I
~
1.0
1.5
Gas V e l o c i t y in t h e C h a r / A S h S e p a r a t o r , v 1 rn/sec 0.15 m
-Yr
-Air Inlet
Figure 3. Concentration of dolomite in the mixture withdrawn from the char-ash separator.
m
loo*
*
Figure 2. Plexiglas model (114.3 mm diameter) with conical bottom. simulated in a 114-mm diameter Plexiglas unit as depicted in Figure 2. The column is similar to that described previously (Chen and Keairns, 1975) except that in this study a conical bottom of 45O or 60' from the horizontal with a 25.4-mm diameter nozzle was used. The nozzle was plugged with a rubber stopper (attached to a 6.35-mm tube which could be moved up and down) before the bed was filled with the char-ash mixture. Air was introduced into the column as the plug was at first slowly and then more readily removed. A timer was started once the nozzle was completely open. Ash particles, which are heavier than char, separated from the mixture by falling through the nozzle countercurrent to the high gas velocity. Lighter char particles were entrained back to the fluidized mixture. For a given period of operation, e.g., 1to 10 min, the separated particles were weighed and screened to determine the rate and degree of separation. These particles were then returned to the fluidized bed for the next run. No provision was made for continuous feeding of simulated ash into the bed. The rate and degree of separation were determined as a function of the gas velocity at the nozzle, the bed composition, the cone angle, and the bed height. Experimental Results and Discussion Low-Velocity Separation. Fluidization and Separation Characteristics. A nonuniform air distribution with a high gas velocity at the center nozzle and low gas velocities a t the conical annular distributor and the annular char-ash separator was introduced to simulate the operating characteristics projected for the Westinghouse gasifier-combustor (Merry et al., 1976). Segregation of particles occurred in the bed when (1)a high solid circulation induced by a high-velocity jet existed between the combustor and the gasifier, and (2) minimally fluidized particles existed in the outer annulus of the gasifier (immediately above the conical distributor) and in the char-ash separator. The jet was stable and confined to the center of the combustor when its velocity was high (>5.8 m/s); a t lower velocities it oscillated and was unstable. Large bubbles were occasionally produced a t the jet tip, resulting in slugging bed operation in the agglomerating combustor for velocities
