Captura de SO2 Con Caliza en LFP
Captura de SO2 Con Caliza en LFP
Captura de SO2 Con Caliza en LFP
Introduction
Since the uncalcined limestone is non-porous, the reaction takes place at the
external surface of the particle to form a product (CaSO4) layer. The rate of reaction has
been investigated mainly by use of thermogravimetric analysis •1-7•. The reaction rate
was observed to be governed by both chemical kinetics and diffusion of SO2 through the
product layer. The reaction rate constant, order of reaction, and effective diffusivity
through the product layer have been evaluated for different types of limestone.
17
THERMAL SCIENCE: Vol. 7 (2003), No. 1, pp. 17-31
PFBC experiments
18
Shimizu, T., et al.: Capture of SO2 by Limestone in a 71 MWe Pressurized Fluidized ...
cross section of the reactor was 7 ´ 4 m and the total height was 8 m. Bed height was
varied with load change to control the number of boiler tubes immersed in the dense
bed, thus to control the heat recovery rate; the bed height was 3.5 and 2 m at the full
load and 50% load, respectively. Total pressure was also varied with load and it was 1.1
and 0.7 MPa at the full load and 50% load, respectively. Temperature in the bed was
fixed at 1100-1135 K. The gas velocity was kept constant at 0.9 m/s. The carry-over size
of limestone was estimated to be 0.25 mm. The gas residence time in the freeboard was
approximately 5-7 s. In the previous study •10•, the results obtained without fly ash
recycle (Phase-1 configuration in fig. 1) was analyzed during combustion of one kind of
coal (BA coal). In the present study the results obtained with fly ash recycle was
analyzed (Phase-2 configuration in fig. 1). The fine particles greater than 0.075 mm
were captured by cyclones and recycled to the bed. The fly ash smaller than 0.075 mm
was captured by ceramic filters after the cyclones and drained from the system. The ex-
perimental conditions and SO2 concentration in the flue gas are summarized in tab.1.
In the present work, four kinds of coal were employed as fuel (tab. 2). The fuel
was mixed with water and limestone to form paste and fed to the bottom of the reactor.
One kind of limestone was employed as sorbent (tab. 3). The size distribution of lime-
stone is shown in fig. 2. Ca/S molar ratios of 2.5-7.7 were adopted for the present study.
Such high Ca/S ratio was necessary not to achieve SO2 capture but to maintain the bed
height during combustion of low-sulfur coal since the bed material, which mainly con-
sisted of limestone, was lost by attrition; certain bed height is required for heat transfer
to the boiler tubes, thus excess limestone feed was required to compensate the loss of
limestone. The excess limestone feed resulted in quite low emission of SO2 (<30 ppm)
as shown in tab. 1.
19
THERMAL SCIENCE: Vol. 7 (2003), No. 1, pp. 17-31
Table 1. Operation conditions, SO2 emission and bed material surface area
Bed
Bed
Coal feed Limestone O2 SO2 material
Power Ca/S material
Run ID Coala rate (dry) feed rate conc.b conc.b surface
•MWe• •–• weight
•kg/s• •kg/s• •%• •ppm• area
•103 kg•
•105 m2•
1998/11/20 BA 66.0 3.6 6.37 0.218 56.1 6.2 13 2.72
1988/12/16 BA 66.0 3.6 6.06 0.210 54.8 6.6 7 2.19
1999/3/18 BA 35.0 4.8 3.68 0.170 38.2 9.5 3 1.37
1999/4/21 NT 37.0 3 3.56 0.209 37.8 7.5 30 1.35
1999/9/17 BA(7) + AD(3) 57.0 7.7 5.73 0.400 49.6 6.8 16 1.71
1999/10/5 DT 32.5 2.5 3.61 0.294 33.0 9.1 29 1.21
1999/10/11 BA 50.8 4.8 5.03 0.232 44.9 7.2 3 1.71
1999/10/18 BA(5) + DR(5) 36.8 2.5 3.83 0.204 35.3 7.1 21 1.79
1999/10/27 BA(5) + DR(5) 36.5 2.5 0.00 0.207 32.9 7.9 13 1.69
a – Coal1(p) + Coal2(q): coal1 and coal2 were mixed with a weight ratio of Coal1:Coal2 = p:q
b – Measured at the outlet of gas turbine
20
Shimizu, T., et al.: Capture of SO2 by Limestone in a 71 MWe Pressurized Fluidized ...
Table 4. Evaluation of the formation rate of fine limestone particles due to attrition
Feed rate Fine Ca
Feed rate Ca flow out
Feed rate of Ca in formation
Fly ash Ca in Ca in coal of Ca in rate in fly
of ash in fine rate by
Run ID drain rate fly ash ash coal ash, ash,
coal limestone, attrition,
•kg/s• •%• •%• FCa,CA FCa, FA
•kg/s• FCa,LF FCa, AT
•mol/s• •mol/s•
•mol/s• •mol/s•
The feed rate of fine limestone particles was calculated from the feed rate of
limestone (tab. 1) and size distribution of limestone (fig. 2). Feed rate of calcium in coal
ash was calculated from the coal feed rate, ash content in the coal and calcium content in
the ash, assuming that all of the ash was broken into small particles and carried over by
the flue gas since the ash was fragile. As shown in tab. 4, the contribution of calcium in
coal ash to total flux of Ca in the fly ash was only minor, thus the assumption of the
behavior of coal ash was not so important. By subtracting FCa,LF and FCa,CA from total flux
of Ca in fly ash (FCa,FA), Ca in the fine particles formed by attrition was obtained as shown in
tab. 4 as follows:
21
THERMAL SCIENCE: Vol. 7 (2003), No. 1, pp. 17-31
FCa, AT M
a= (3)
rA
External surface area of the bed material (A) was calculated from the total mass
of the bed material (WBM) and size distribution of the bed material (fig. 3) as follows:
6WBM w
A= å i (4)
r D pi
where r, Dpi, and wi are density of limestone, particle size of bed material of i-th fraction,
and mass fraction of bed material that has size of Dpi, respectively. The mass of the bed
material was estimated from the pressure drop across the bed. The amount of the bed
material and external surface area of limestone are shown in tab. 1.
Figure 4 shows the relationship between plant power output and attrition rate
for various coals. The average attrition rate was approximately 1 mm/h and the influence
of plant power output on the attrition rate was only minor. This is attributable to the
constant gas velocity throughout the operating conditions; the gas velocity was kept
constant by changing both air feed rate and total pressure in the vessel. Also the fuel type
had little influence on the attrition rate. This indicates that the interaction between the
fuel and the limestone was not important.
Concentration of SO2 in the flue gas was between 3 and 30 ppm (tab. 1). Figure
5 shows the relationship between Ca/S molar ratio and SO2 removal efficiency. Ca/S
ratio was calculated from the feed rate of limestone, feed rate of coal and sulfur content
in the coal. SO2 removal efficiency was calculated from the sulfur content of the fuel,
coal feed rate, air feed rate and the concentration of SO2 in the flue gas. Though it is
usual that the SO2 emission is discussed in relation to Ca/S ratio, Ca/S ratio was not a
good index to describe the sulfur capture behavior in the present PFBC.
22
Shimizu, T., et al.: Capture of SO2 by Limestone in a 71 MWe Pressurized Fluidized ...
Figure 4. Effect of plant power output and Figure 5. Effect of Ca/S ratio on SO2 removal
fuel type on attrition rate efficiency
23
THERMAL SCIENCE: Vol. 7 (2003), No. 1, pp. 17-31
the direction to the particle’s center. The position of the center of particle is fixed. Xe
and Xc denote the positions of external surface and unreacted core surface at time = t,
respectively. Product layer thickness d is:
d = Xc - Xe (5)
In the previous study, the overall reaction rate was found to be mainly governed
by the diffusion of SO2 through the product layer. Thus the assumption of “diffusion
controlling” was employed. The specific reaction rate (SO2 capture rate per unit external
surface area), qS, under diffusion controlling condition is given as follows:
De C
qS = (6)
d
where De and C are the effective diffusivity through the product layer and concentration
of SO2 at the external surface, respectively. A value of De = 1.5×10-9 m2/s was adopted;
this value was based on the results of the TGA study by Qui and Lindqvist •7• and
modified by Shimizu et al. •9• for the application to thin product layer.
The change in the distance between the unreacted core surface and the initial
particle surface is given as follows:
dX c D e CM
= (7)
dt rd
where M and r are molecular weight of CaCO3 and density of limestone, respectively.
The external surface is removed by attrition. Figure 7 illustrates the change in
the particle radius with time under intermittent solid attrition conditions. To make the
model simple, periodical attrition was assumed. The external surface was removed with
an interval of t, thus the change in Xe was given as follows:
24
Shimizu, T., et al.: Capture of SO2 by Limestone in a 71 MWe Pressurized Fluidized ...
dd = D e CM (9)
( jt < t < ( j + 1)t, j = 0, 1, 2,... )
dt rd
If the fresh CaCO3 surface appears when attrition occurs, the thickness of the
product layer at the moment of attrition t = jt (j = 0, 1, 2, …) is zero. By solving eq. 9, the
product layer thickness is given as a function of time as follows:
12
é 2D CM (t - jt) ù
d(t) = ê e ú ( jt < t < ( j + 1)t, j = 0, 1, 2,... ) (10)
ë r û
The average rate of increase in the product layer thickness during one period of
attrition, from t = 0 to t = t, is given as follows:
12
dd æ 2D e CM
= d(t) = ç
ö
÷÷ (11)
dt average t ç rt
è ø
Thus the average SO2 capture rate per unit surface area, rS, is given as follows:
12
2D e r ö
rS = d(t) r = æç ÷ C1 2 (12)
t M è Mt ø
The criteria if the fresh CaCO3 surface appears when intermittent attrition
occurs is given as follows:
12
æ 2D e CMt ö
at > d(t) = ç ÷÷ (13)
ç r
è ø
The conversion of CaCO3 to CaSO4 in the fragment is given as the ratio of
thickness of the product layer to the removed thickness by attrition:
12
æ 2D e CMt ö
çç ÷÷
h= d(t ) = è r ø (14)
at at
In the present work, the fresh CaCO3 surface was assumed to appear when
intermittent attrition occurs since the overall attrition rate was far higher than that of
overall SO2 capture rate. The fragment formed by attrition contained unreacted CaCO3,
i. e. h < 1. Thus at is considered to be greater than d(t) and fresh CaCO3 appears.
25
THERMAL SCIENCE: Vol. 7 (2003), No. 1, pp. 17-31
Formation of SO2 from fuel occurs during volatile matter combustion as well as
char combustion. Volatile matter combustion is assumed to take place at the bottom of
the bed; Suzuki conducted bench-scale PFBC experiments using a transparent quartz
reactor installed in a pressure vessel and found that combustion of volatile matter took
place only at the bottom of the reactor •12•. Thus SO2 concentration at the bottom of the
combustor, C(0), is given by the rate of SO2 formation from volatile matter and gas flow
rate, VG:
G s VM coal
C(0) = coal coal (15)
VG
where Gcoal and scoal are coal feed rate and sulfur content in fuel, respectively. VMcoal is
the portion of sulfur released as volatile matter. In the present study, VMcoal is assumed
to be the same as volatile matter content of fuel.
SO2 formation from char is assumed to occur uniformly throughout the bed
since the solids are considered to be completely mixed in the fluidized bed. The SO2
evolution rate per unit mass of bed material, RF, is given as follows:
In the 71MWe PFBC, the particle size in the bed was mainly 0.25-5 mm (fig. 3).
For fluidized beds consists of such coarse particles, in which gas velocity is higher than
bubble rising velocity, mass transfer resistance between bubble and emulsion is not so
important, thus the bed can be treated as a plug-flow reactor •13•. In such reactor, the
change in SO2 concentration with contact with solids is given as follows:
VG dC
= RF - RS (18)
dWBM
26
Shimizu, T., et al.: Capture of SO2 by Limestone in a 71 MWe Pressurized Fluidized ...
By solving this equation numerically with initial value C(0) from eq. 15, the
concentration profile along the bed height as well as flue gas SO2 concentration were
obtained.
For only DR coal, the deviation between the model and the experimental result
was remarkable. This deviation is attributable to the limitation of the present simple
reaction model. The present model assumes that the fresh CaCO3 is exposed when
attrition occurs (eq. 13). This condition is rewritten in terms of solid utilization as
follows:
h = d(t) < 1 (19)
at
Due to the high sulfur content, the condition in the reactor for DR coal is close
to the limit. As shown in fig. 9, the total SO2 capture rate per total attrition rate for DR
27
THERMAL SCIENCE: Vol. 7 (2003), No. 1, pp. 17-31
coal was the highest among fuels tested. Total SO2 capture rate within the reactor (Y)
was given from the feed rate of S in the coal and the flux of SO2 in the flue gas as follows:
hexp er = Y (21)
FCa, AT
For DR coal, the value of hexper was nearly 0.8 and those for others were less than 0.6.
Though in the present model only single value of t was assumed, it is conceivable that t
has distribution. When the experimental condition is close to the limit given as eq. 13 or
eq. 19, some limestone particles may exceed the limit if t has distribution. When the utili-
zation is sufficiently smaller than unity, all
the particles fall in the limit even if t has dis-
tribution.
Figure 10 shows the relation between
sulfur content of fuel and limestone utiliza-
tion efficiency calculated theoretically ac-
cording to eq. 14. Similar relationship be-
tween sulfur content and limestone utiliza-
tion efficiency to fig. 9 was obtained. The
present model was effective not only estima-
tion of SO2 emission but also limestone utili-
zation efficiency.
As discussed above, the attrition interval
Figure 9. Effect of sulfur content of fuel on (t) plays an important role in determining
the ratio of total SO2 capture rate in PFBC desulfurization behavior of limestone. How-
and overall limestone attrition rate ever, the period of attrition in PFBCs has
not yet been evaluated. The attrition inter-
val can be evaluated if the size distribution
of limestone fines formed by attrition is
obtaiend; the size of limestone fragment is
considered to be approximately at. How-
ever, the fly ash includes both limestone
fines and coal ash, thus only the Carich par-
ticles should be picked up and the size of
such particles should be measured. For such
purpose, CC-SEM (Computer-Controlled-
-SEM) may give useful information. How-
ever, the size distribution data is not yet
Figure 10. Effect of sulfur content of fuel on available at this moment. This is a subject of
the utilization of limestone (theoretical future works.
results)
28
Shimizu, T., et al.: Capture of SO2 by Limestone in a 71 MWe Pressurized Fluidized ...
Conclusion
A wide variety of coals were burnt in a 71 MWe pressurized fluidized bed com-
bustor under fly ash recycle conditions. The emission of SO2 was between 3 to 30 ppm.
SO2 removal efficiency was not correlated with Ca/S ratio. Attrition rate of limestone
was evaluated from the drain rate of calcium in fly ash. Limestone attrition rate was ap-
proximately 1 mm/h and it was not affected by the coal type and plant power output. The
ratio of total sulfur capture rate to limestone attrition rate was nearly proportional to the
sulfur content of fuel.
A simplified model of SO2 capture by limestone under pressurized fluidized
bed combustion conditions was applied to the analysis of the present PFBC. In this
model, intermittent attrition of limestone was assumed. By giving a value of the period of
attrition of 5 hours, which was obtained in the previous work without fly ash recycle, the
experimental results agreed well with the model for most of the coal, except for high sul-
fur content coal.
Acknowledgement
Nomenclature
29
THERMAL SCIENCE: Vol. 7 (2003), No. 1, pp. 17-31
Greek letters
References
•1• Snow, M. J. H., Longwell, J. P., Sarofim, A. F., Direct Sulfation of Calcium Carbonate, Ind.
Eng. Chem. Res., 27 (1988), 2, pp. 268-273
•2• Hajaligol, M. R., Longwell, J. P., Sarofim, A. F., Analysis and Modeling of the Direct Sulfation
of CaCO3, Ind. Eng. Chem. Res., 27 (1988), 12, pp. 2203-2210
•3• Iisa, K., Hupa, M., Rate-Limiting Processes for the Desulphurisation Reaction at Elevated
Pressures, J. Inst. Energy, 65 (1992), 4, pp. 201-205
•4• Krishnan, S.V, Sotirchos, S.V., Sulfation of High Purity Limestone under Simulated PFBC
Conditions, Canadian J. Chem. Eng., 71 (1993), 2, pp. 244-255
•5• Krishnan, S.V, Sotirchos, S.V., A Variable Diffusivity Shrinking-Core Model and its
Application to the Direct Sulfation of Limestone, Canadian J. Chem. Eng., 71 (1993), 5, pp.
734-745
•6• Zevenhoven, R., Yrjas, P., Hupa, M., Sulfur Dioxide Capture under PFBC Conditions: the
Influence of Sorbent Particle Structure, Fuel, 77 (1998), 4, pp. 285-292
•7• Qui, K. Lindqvist, O., Direct Sulfation of Limestone at Elevated Pressures, Chem. Eng. Sci.,
55 (2000),16, pp. 3091-3100
•8• Sakuno, S., Shimizu, T., Misawa, N., Suzuki, N., Ueda, H., Sasatsu, H., Gotou, H., SO2
Emission Control by Limestone in a 71 MWe pressurized Fluidized Bed Combustor, Part 1.
Attrition and Fragmentation of Limestone (in Japanese, Nihon-Energy-Gakkai-Shi (J. Jpn.
Inst. Energy), 80 (2001), 8, pp. 747-757 )
•9• Shimizu, T., Peglow, M., Sakuno, S., Misawa, N., Suzuki, N., Ueda, H., Sasatsu, H., Gotou,
H., Effect of Attrition on SO2 Capture by Limestone under Pressurized Fluidized Bed
Combustion Conditions – Comparison between a Mathematical Model of SO2 Capture by
Single Limestone Particle under Attrition Condition and SO2 Capture in a Large-scale
PFBC, Chem. Eng. Sci., 56 (2001), 23, pp. 6719-6728
•10• Shimizu, T., Peglow, M., Yamagiwa, K., Tanaka, M., Sakuno, S., Misawa, N., Suzuki, N.,
Ueda, H., Sasatsu, H., Gotou, H., A Simplified Model of SO2 Capture by Limestone in a 71
MWe Pressurized Fluidized Bed Combustor, Chem. Eng. Sci., 57 (2002),19, pp. 4117-4128
30
Shimizu, T., et al.: Capture of SO2 by Limestone in a 71 MWe Pressurized Fluidized ...
•11• Shimizu, T., Peglow, M., Yamagiwa, K., Tanaka, M., Comparison amoung Attri-
tion-Reaction Model of SO2 Capture by Uncalcined Limestone under Pressurized Fluidized
Bed Combustion Conditions, Chem. Eng. Sci., (in press)
•12• Suzuki, Y., Fundamental Studies on Pressurized Fluidized Bed Combustion at NIRE (in
Japanese), Seminar on Practical Use of Pressurized Fluidized Bed Combustion System
December 6, 2000, Tokyo, Soc. Chem. Engrs. Japan, 2000, pp. 1-10
•13• Kunii, D., Levenspiel, O., Fluidization Engineering (2nd edition), Butterworth- Heinemann,
Stoneham, USA, 1991, Chapter 12
Authors’ addresses:
M. Peglow
Otto-von-Guericke University Magdeburg
Universitätplatz 2,
D-39016 Magdeburg, Germany
31