-------
£ «
o>
c
a
«i _ _
01 G> ^" O CO
Jf Z O >-
ma a B a
m a
00
ID
m VCD a
BO
O
B
B B
C3 <=
-t->
c
O!
O
O
E
•r-
ID
O
O
o
c
M-
CO
V)
(O
CO
s_ .
d) (A)
E *<0
O O
O O
CM O)
O 4J
4-> cn
OJ i^
o
s- s-
O) O
Q- M-r
O)
CD
UOIS-I3AUO3 2QS
38
-------
3.3 PERCENT SULFUR IN THE COAL
SOg emissions obviously increase with increased sulfur content of the
coal. But Figure 7 shows that the percent conversion of fuel sulfur to SC"2
(dashed line) also increases from approximately 50 to 100 percent when sulfur
.content increases from 0.5 to 1.5 percent. Beyond 1.5 percent sulfur in the
coal the conversion remains constant at approximately 100 percent. The rea-
son for this increase in percentage conversion appears to be due mostly to
the change in coal characteristics as the sulfur is increased. The low sul-
fur content coals represent subbituminous and lignitic coals, while the
higher sulfur content coals represent bituminous coals. Together with a
reduction in sulfur content, ash properties also change, causing the reduc-
tion in percent conversion to SOp.
The three solid lines indicate the allowable percent conversion of
the fuel sulfur in order to maintain S02 emissions at the 516 ng/J (1.2
Ib/MBtu) level promulgated by the NSPS without any scrubbing device. For
example, a boiler firing bituminous coal with a typical heating value of
30,238 J/g (13,000 Btu/lb) and a sulfur content of 2.0 percent would have
to retain at least 60 percent of the sulfur to comply with the NSPS without
an added control (see short dashed line). All subbituminous coal-fired
boilers investigated fell below their curve indicating that no SO^ control
would be necessary to meet the federal standards. Emissions from lignite-
fired boilers were slightly above the allowable limit. S02 emissions from
bituminous-fired boilers were far above the federal standards, indicating
that SOp control devices .would be necessary to meet the 516 ng/J level. It
should be noted that the three solid lines represent typical coals with
typical heating values. The heating values chosen to calculate these curves
are not necessarily the heating values of the coals used in the reported
field tests, but represent a good approximation for each generic coal type.
3.4 BURNER STOICHIOMETRY
One of the mechanisms by which S03 can be formed is the S02-atomic
oxygen reaction. Based on this theory, an increase in S03 production should
be observed when the percent burner stoichiometry (excess air) is increased.
This increase in SO,, production would shift the S02-S03 equilibrium composi-
tion toward SOg, thus reducing S02 emissions. Figure 8 shows the percentage
conversion of sulfur to S02 as a function of burner stoichiometry for the
39
-------
in
o
CO
CM
O
co
o
-p
OJ
o
o
to
o
o
q-
LU
O)
suna
UB ao^ paSeaaAB (^uaoaad) uoisJ3Auoo
40
-------
.
s
CM
O
130
120
110
100
90
80
70
60
A Huntington N = 2 tangential
O Unit "B" front wall
@ Assumed bituminous coal based
on fuel analysis Widows Creek
No. 5 front wall
X Unit "A" vertical
0 Unit "Du H.O.
© Unit "C" tangential
^ Berry No. 4 tangential
© Berry No. 2 tangential
wf
^7
V
3
e
#©
u ^^ A A
e e e
x
x_
e
e
100 HO 120 130 140 150
Percent burner stoichioraetry
160
Figure 8. Effect of burner stoichiometry on the percentage
conversion of bituminous coal sulfur to SC•
41
-------
bituminous coal and Figure 9 for all the subbituminous coal data. No trends
of S02 reduction with increased furnace excess air can be seen. The data
are scattered to such a degree that no clear trend can be seen whatsoever, j
even within each boiler test run series.
Figure 10 and 11 show the percent conversion of coal sulfur to S03 :
for three boilers. An increase in the percentage SOs emissions can be seen
for Unit "D" boiler, although the increase is rather speculative since it is ,
based on only few data points. Figure 11 shows S03 conversion for a
vertically-fired boiler. The percent S03 to total sulfur oxides was higher
than for Units "B," "C," and "D," ranging from 1.6 to 9.2. However, S03 :.
seems to be insensitive to changes in burner stoichiometry. Additional data i
are necessary to draw any conclusions on the higher S03 percentages in
vertically-fired units than other boiler firing types. ; ;
3.5 BOILER FIRING RATE
The equilibrium mixture of S02 and SOs is both a function of tempera- •
ture and pressure. Lowering the temperature shifts the equilibrium toward •
S03 production. Thus, it would be expected that as the boiler firing rate is|
reduced, and lower gas temperatures occur in the firebox, an increase in S03
emissions would take place with a consequent decrease in S02 emissions. ;
Figures 12 and 13 show that this may not be the case. Again, there is con-
siderable scatter of both the individual data and the effects of firing rate '
changes on different boilers. Although the available data are insufficient
to justify any conclusion about conversions as a function of firing rates
for individual boiler types or coals, it is clear that there is no general \
trend for all boilers and coals.
i
3.6 BOILER SIZE
Another boiler parameter that could affect sulfur conversion is unit !
capacity. To assess this posssibility, emission rates (Mg/hr) of S02 were i
plotted as a function of boiler size with sulfur content of the coal indi- .
cated for each point (Figure 14). As expected S02 emissions increase with
both boiler size and sulfur content of the coal. For comparison the emis- [
sion limit stipulated by the current NSPS is also shown. All points below i
this standard represent low sulfur western and lignitic coals. '
42 [
-------
120
110
*« 100
cu
E
o
cj
o
co
90
80
70
100
Columbia No. 1 tangential
110 120 130 140
Percent burner stoichiometry
CO
-------
2.5
® Unit "B" (FW)
-------
10
O)
en
ra
0) ,
S 6
0)
Q.
CM 4
O
-------
0
Huntinton (T)
120
u
10
03
•o TOO
Ol
en
K)
!_
-------
no
100
Columbia No. 1 Q Wm. J. Neal (FW)
(T) g) Leland Olds (HO)
A Comanche #1 (T)Q Hoot Lake (T)
A Station #1 (T) Q3 Milton Young (Cy)
A Station #2 (T) ® Station #3 (Cy)
A Subbituminous coal
O Lignitic coal
A
AA
90
CD
1 80
£
OJ
n
B
I
o
u
o
oo
70
60
s
40
50 60 70 80 90
Percent firing rate (MCR)
100
no
Figure 13. Effect of firing rate on SOg emissions for
lignite and subbituminous coal-fired boilers,
47
-------
O Bituminous
A SubbHumlnous
D Lignite
Numbers above symbols indicate
percent sulfur in the coal.
3 4
Boiler MCR x 1012 J/h
Figure 14. Effect of boiler size on S02 emission rate,
48
-------
To determine whether sulfur conversion does, in fact, depend on
boiler size, one needs to analyze the sulfur emission rate per unit energy
produced (MW-hr). If such a dependence exists, it is probably the indirect
consequence of differences in boiler efficiency (for a given fuel). To
check this possibility, mass emission rates per energy output (kg/MW-hr)
were plotted as a function of boiler size (see Figure 15(a)). The results
show considerable scatter with no apparent correlation to boiler size. The
major cause of this scatter is, of course, the variation in coal type, sul-
fur and moisture content, and heating value among the data points. Coal
type and sulfur content have already been shown to affect sulfur emissions.
Variations in coal moisture content and heating value effect emissions when
measured in mass per energy output because of their effects on boiler
efficiency. One can try to separate out the effects of coal type and sulfur
content fay (1) plotting the data for each coal on a different graph (Figures
15(b) to 15(d)) and (2) comparing emissions from different sized boilers when
each fires coal of approximately the same sulfur content as the others.
The following selected examples show that there is no unique relationship
between emissions (per energy output) and boiler size, even for a given coal
type and sulfur content.
9 Bituminous: coals with S = 2.60 ± 0.04 percent were fired in a
20-, 125-, and 350-MW boiler. Emissions per energy output in-
creased with size from about 6.2 to 18.4 kg/MW-hr. However, a
1.45-percent S coal fired in a 125-MW boiler emitted at essen-
tially the same rate as did a 270-MW boiler burning a 1.5-percent
S coal and another 270-MW unit using a 1.2-percent S coal.
9 Subbituminous: an 0.52-percent S coal in a 425-MW boiler emitted
less than a 350-MW unit firing 0.49-percent S, but about the same
as one would expect the other two units (330 MW and 350 MW) to
emit if they were firing 0.5-percent S coal (instead of 0.61 and
0.72-percent S)
@ Lignite: a 20 MW-boiler emitted more burning 0.64-percent S coal
than did a 215-MW boiler burning 0.77-percent S coal, but a 50-
MW unit burning 1.17-percent S coal emitted substantially less
than did a 250-MW boiler on 1.3-percent S coal
49
-------
Q Bituminous
^ Subbituminous
Q Lignite
Numbers above symbols indicate
percent sulfur in the coal.
, Avg. S02 emissions:
^kg/MW-hr
20
15
1.55
10
2TT -
o Q
b9
0.77
2.56
O
0.61
0.52
•t
100
200 300 400
Boiler size, WW
500
Figure 15(a). Effect of boiler size on S02 emissions.
50
-------
20
15
1 10
01
CSI
o
1/1
en
S_
O)
2-6
3.47
O
2.64
O
2.9
O
2.07
O
1.45
0
5
gl-
Numbers above symbols indicate
percent sulfur in the coal.
2.56
TOO
200 300
Boiler size, ("
400
500
tj
-------
20
«= 15
Numbers above symbols indicate
percent sulfur in the coal.
10
e
-------
20
15
1.55
D
10
01
en
0.64
1.17
D
100
Numbers above symbols indicate
percent sulfur in the coal.
1.3
1.0
O
5
CD
<
0.77
Q
300
Boiler Size, MW
400
500
Figure 15(d). Effect of boiler size for lignite coals.
53
-------
These relationships are seen more clearly in Figure 16 (a qualitatively de-
termined "best-fit" straight line has been added as a visual aid; it cannot
be used too rigorously because a straight line relationship between emis-
sions and sulfur content is valid only for "constant" coal and boiler
efficiency). In some cases, the smaller units appear above the larger ones
(for a given sulfur content and coal type), whereas in other cases, they
fall below.
One reason there may be no clear relation between size and sulfur
emissions is that boiler unit efficiencies do not vary much with size for
boilers larger than 100 MW (Reference 16). The variation in efficiency is
typically only between 87 and 90 percent. Even this variation is probably
due more to age than size, because the larger units tend to be the newer
ones. With the-current trend toward the installation of medium-sized
boilers rather than the very large ones, this dependence of efficiency on
size will diminish.
Figure 16 also suggests that the lignitic coals cause higher S02
emission rates, when referenced to energy output, than do the bituminous
coals with the same sulfur content. No comparisons can be drawn with sub-
bituminous coals, however, because of a lack of data.
Unlike S02 emission rates, which depend directly on fuel sulfur con-
tent, fuel heating value, boiler firing rate, and possibly boiler firing
configuration, S02 conversion rates do not necessarily depend on boiler type/
size. It was shown in Figure 7, however, that S02 conversion increases with
fuel sulfur content up to 1.5 percent because of the change in coal charac-
teristics. To see if this effect carries over when emissions are related to
boiler size, Figure 17 was prepared. This plot shows S02 conversion as a
function of boiler output size. Examination of the figure suggests that S02
conversion increases somewhat for the bituminous coals as boiler size in-
creases. A similar conclusion appears to hold for lignite. The strongest
correlation, however, still is with sulfur content and type of coal. A much
larger data base would be required to more rigorously evaluate the depen-
dence of S0£ conversion on boiler size.
54
-------
20
s_
HI
Q Bituminous
^ Subbituminous
Q Lignite
Number above symbols indicate
unit size in MW.
1.5 2.0
Percent S in coal
Figure 16. Effect of sulfur content on S02 emissions.
55
-------
2.56
0.75
A
100
90
•§
en
IB
1 80
IO
*?
conversion
»j
o
CM
%
6
c
2.6 <•>
0 0.49
A
'#, W
ys o
1.0 1.5
EJ O
2.64 W
0.64 0 0.77
Q uJ
3.5
.1.17 0 0.55
EJ O
0.72
A
0.61
A
100
200
300 400 500
Boiler design loading (MW)
O Bituminous
G3 Lignite
A Subbituminous
Numbers above symbols indicate
the average sulfur content of
the coal.
600
700
0.55
A
800
Figure 17. Effect of boiler size on S02 emissions.
56
-------
SECTION 4
GASEOUS SULFUR EMISSION ACROSS PARTICULATE COLLECTION DEVICES
Virtually all coal-fired power plants are equipped with particulate
control devices to capture the flyash they emit. From the perspective of
SO control, therefore, the typical boiler — the so-called "uncontrolled"
X
unit - is one with particulate controls. Since the data reported in Section
3 were measured in the ducts.ahead of any control devices, the actual SO
A
emissions from the plant could be different than the measured values. To
determine whether this is true, in fact, data were collected for SO emis-
A
sion rates on both sides of particulate control devices; these data are
reported here.
The results of four series of tests (Reference 6) in which sulfur
oxides were measured at both the inlet and the outlet of particulate col-
lection devices are summarized in Tables 6(a) and 6(b). Each test series
is for a different power plant and control system. Two boilers (Units "A"
and "C") have a cyclone followed by an electrostatic precipitator (ESP),
one (Unit "B") has only an ESP,, and the last one (Unit "D") has only a
cyclone.
It is interesting to note that the average inlet S02 mass loadings
for three of the units was nearly the same. On the average, the S0? mass
loading across the collection devices decreased slightly for units "C" and
"0", however, they increased significantly for unit "B". S0? emissions are
not expected to change significantly across these collection devices. Large
differences in emissions across these collectors can be attributed more to
measurement errors than effects of the collectors.
57
-------
In the case of sulfur trioxide emissions, units "B", "C", and "D"
had nearly the same concentration at the inlet, but unit "A" produced a
considerably larger quantity of SO,. The exit streams for all four collec-
tor devices had similar concentrations of SO.,. As a result, the concentra-
tion of SO, across the mechanical dust collector-electrostatic precipitator
for unit "A" was greatly reduced, while the other three units showed slight
increases. These small increases are probably within the uncertainty of
the measurement techniques, therefore it is difficult to identify trends.
Evidently the collecting devices for unit "A" were successful in removing
some SO, from the flue gas. This substantial reduction could be due to
leakage and temperature decrease of the gas stream across the collectors.
The cooling of the flue gas could have resulted in condensation of the SO-
and formation of sulfuric acid mist. The resulting mist as well as some
sulfur trioxide gas could be adsorbed on the flyash particulates. Then
upon 'removal of these particles, the concentration of SO- would be reduced.
In addition, for the case of an electrostatic precipitator the acid mist
particles could be ionized and collected in the precipitator.
In conclusion, the data do not show any trends. SO- emission de-
creased in two cases (by 6.5 percent on the average) and increased in two
other cases, where similar collection devices were used (by 24 percent on
the average). In one case with relatively high S03 emissions, the combina-
tion of an ESP and mechanical collector removed over 80 percent of the SOg.
Inconsistencies in the S02 emission data across particulate collection
devices can be attributed to the measuring technique used. These techniques
consisted of single point grab samples from large and split ducts. A dis-
cussion of measurement techniques is presented in Appendix C.
58
-------
86S-1
(0
oo
UJ
CJ)
Q
O
O
LU
O
O
•<
o
^3-
§ =
Sj|,|
=>~
«1I
Percent
Conversion
0>
si's
^s-5-
z
Boiler - Unit A vertically
fired. High volatile
bituminous coal.
co CD co r*. co o
i— o in o co co
co o co cn *a- co
+ + +1 ' "*""*"
co o i— «* i— r-~
CD CD i— CD .— O
ro o «* in •— i—
**" "~ r- r-
co cc cn o cn CD
cn <=r O r^ "3- CD
o =• in r*. in r^
LO .- cn •* CM •*
10 CO CO i— tO r—
'
» » cv, «,
in ua en **• o m
r~- n j — cn tf} in
cr.
«
Mechanical dust collector
1n series with electro-
static precipitator
Boiler -Unit B front wall
fired bituminous coal.
Original data given in
lb/1000 ft3 of flue gas.
Refer to Appendix A for
conversion testers.
o o* o ro
o o r^ o co
5 f V ". +
VO CO CO CO i —
S S C S f
in
00000
CO
r-. ys o «3 to
CO f- CD ^ CM
•— m CM co •*
CTI cn CM <^- i —
O CD r- CD CD
oo ~ » a o.
CM
to CO i— CD f—
CD r-^ CO O ^~
co CM co r— 10
cn
"C
Electrostatic precipitator
Boiler - Unit C tangentially
fired. High volatile
bituminous coal.
CD CO CD O Ol
ro i r^ CD in 5-
•" + - *
r*. o co i— i — CM
1 1 1 + CM 1
1
O •— r— CO 1— (—
CM cn r^ cn CM CM
co *a- ko \n co
co 10 co cn in cn
CD 01 cn co o cn
•T CM ^— to VO CD
cn
<
Cyclone separator 1n
series with electrostatic
precipitator
Boiler -Unit D horizontally
opposed. High volatile
bituminous coal. Flyash
reinjection.
O O CM O
in O CM ^ CM
CM 1 «3- i—
CM f— cn cn co
o CM co r— r^.
i i
r- -g- co to
o o o o
r. * CO ^
ro r-^ cn to CM
cn cn co co cn
VO CM *3- CO CO
to CO *3" CD ^T
CD CD O O O
to «* ro «y CM
—
»a- CD co 03 o
CD o cn cn CD
co *a- o co vo
0)
<
Cyclone separator
o
-------
•a
0)
•a
o
o
o
.a
UJ
ca
-=c
6965 -i
IT J^
.
f ' » t" ft t. ; r» * «e OJ
; . SSfG>*«->cser>ocno
i O- ^3
o
tr^^j*"!^ *^ . ^J °i
« * | 2: g ^' "*
' -> •— ,
• ; r^^ ':
; rxS g i — r~ *r o <-> u>
, . ° |-| i "' - CT » ~ "•
' ' 0" '& '
-fg-si s a 2 s s
' 12:3— ,.000000
Boiler - Unit B front wall
fired bituminous coal.
Original data given in
lb/1000 ft3 of flue gas.
Refer to Appendix A for
conversion testers.
o en o fO
0 o r^ o co
i — «3 r~ Lf) CO
IO CO CO
CO CO CM
r-^ r-^ CJ
,— 0 r-
CO I— CM
O ^ O
O CJ O
Boiler - Unit D horizontally
opposed. High volatile
bituminous coal. Flyash
reinjection.
o o co o
in o CM i CM
CM 1 "31 •— •
co •— cn en co
O CO CO £2 T"
T ' T
t~~ *r co «J
0* 0* O 0
r*. r^ r^. •*
s s a i s
CD O O O
^- co en r~~ o
co r-» cn 10 'CJ
cn cn co co en
co «d- r- co «*•
«J CM CO tf) ^-~«^>,-~"^. .,
1 S i -
,£• «£g
£ ; feu
t Q. ! <-> 0
^2 5; .- ^
F - '^
_j 'C C —
« 'EI'^
uj *_> M a
r- <• :
-3- CD CC CO O
" "• "_^
§ S g S R1
^r <• co CM co ;
J
=.
; •=
!
'
!
I s-
o
11 S- "
e.
. o
^
60
-------
SECTION 5
CONCLUSIONS AND RECOMMENDATIONS
5.1 CONCLUSIONS
The most important result documented by this survey is that the con-
version of sulfur in the coal to S02 emissions depends more on the coal type
and its ash characteristics than on any boiler or design variable considered.
Specific findings are listed below:
1. Sulfur conversion to S02 ranged from 86 to 108 percent for
bituminous coals, from 54 to 114 percent for subbituminous coals,
and from 69 to 97 percent for lignitic coals.
2. Excess air in the furnace and percent firing rate did not seem
to control the conversion of coal sulfur to S0?.
3. The mass emission rate of S02 per energy output (g/MW-hr) does
not appear to depend on boiler size; S02 conversion, however,
does seem to increase slightly with boiler size for bituminous
and lignitic coals.
4. The percent sodium in the coal ash has a significant effect on
sulfur retention in the boiler ash. This is a very important
parameter since the more sulfur retained in the ash, the less
gaseous S02 leaves the boiler. The conversion of sulfur to S09
was reduced from approximately 85 to 50 percent when the sodium
content was increased from 0.9 to 9 percent by weight in a lignitic
coal. Of course, this high sodium content of the ash causes boiler
tube fouling.
5. Cyclone boilers retained the least amount of sulfur in the ash when
burning lignite. Therefore, the S02 emissions from the cyclone
61
-------
boilers burning lignite are generally higher than those from
other lignite-fired boilers with different burner configurations.
6. Gaseous S03 emissions were higher for the vertically-fired boiler
than from any other boiler. However, this trend is not defini-
tive since more data would have to be analyzed to make this result
conclusive.
7. The gaseous SO- content of flue gases is minimum for lignitic
coals due to the formation of sulfate particulates. Gaseous
SOo emission are about the same for bituminous and subbituminous
o
coals.
5.2 RECOMMENDATIONS FOR FURTHER INVESTIGATIONS
The data compiled in this report give some very interesting results
for coal sulfur conversion to S02, S03, and sulfates. However,,in the short
time allowed for this project, all the available data could not be obtained
rapidly enough to allow us to conduct a more detailed and in-depth analysis
of the effect of boiler design and process variables on the emissions. A
substantial amount of additional data was identified and requested, but not
received by the completion data of this task. Sources of the data were
contacted to evaluate the quality and usefulness of their data. They are
listed in Table 7 along with the estimated quantity of sulfur emission data
they could provide.
In addition to the analysis of more emission data, the quality of the
data should be analyzed in more detail to attempt to explain some of the
scatter in the results.
A preliminary investigation was conducted of the sampling techniques
and instrumentation used to collect the data presented in this report (see
Appendix C). Unfortunately the information was too qualitative to identify
sources of error and quantify measurement uncertainties.
62
-------
TABLE T. ADDITIONAL DATA SOURCES
- No. of Emissions
bource Boilers Reported
1 . Sel ker et al . 1 SOg
Reference 11
2. Hollinden 1 ' S02
et al.
Reference 6
3. York Research 10-30 SOX
Corp.
4. Pennsylvania 3 Total
Electric Co. S Balance
5. Oak Ridge 1 NA
National Labs
Report No.
ORNL-NSF-EP-43
6. APCA-1974 1 NA
67th Meeting
7. Mitre "Baseline 1 NA
Measurements"
Test Results
for Cat-Ox
Demonstration
Program ;
TeNs°t R°uns Remarks
28 Low NOX data - Boiler
fired with over fire
air and burners out
of service. Sub-
stoichiometric con-
ditions in the fur-
nace. (Tangential
furnace)
40 Low NOX data. Boiler
fired with burners
out of service.
Substoichiometric
conditions in the
furnace . a
NA Need 2-4 weeks of
work to retrieve the
data.b
NA Some data have been
sent to Aerotherm,
but not in time to
be included in this
report. The remain-
der of the data will
be sent when final
results are ob-
tained, c
NA Identified as con-
taining sulfur
emission data from
pulverized coal-
fired power plants.
Ordered through
Aerotherm library.
NA
NA
o
o
•3-
1
3Nei1 D. Moore of Power Research Staff at TVA has sent (June 1, 1977) fuel
analysis data for their tests on Widows Creek No. 5 conducted in 1974-75.
bHr. B. Epstein of York was contacted (May 16, 1977) in order to obtain data.
York would be willing to send these data to Aerotherm only if York were
reimbursed for the large amount of time they claim it would take to collect
the test data and obtain permission to release them.
telephoned Mr. D. Fyock, Director of Environmental Affairs, Pennsylvania
Electric Company on May 20, 1977, to request their data. Followed telephone
call by a letter.
63
-------
-------
REFERENCES
1. J. A. Cavallero, M. T. Johnston, A. W. Dembrouck, "Sulfur Reduction
Potential of the Coals of the United States, a Revision of Report of
Investigation 7633," Bureau of Mines Report of Investigation 8118, 1976.
2. J. F. Kircher, A. A. Putnam, D. A. Bull, H. H. Krause, J. M. Genca,
R. W. Coutent, J. 0. Y. Wendt & R. Leag, "A Survey of Sulfate, Nitrate
and Acid Aerosol Emissions and Their Control,!1 Battelle Columbus Labora- '
tory Draft Report for W. Steve Lanier of EPA-RTP, September 1976.
3. Homlya, et al., "A Characterization of the Gaseous Sulfur Emissions from
Coal and Oil-Fired Boilers," Fourth National Conference on Energy and
Environment, October 1976, Cine., Ohio.
4. Richard L. Davison, David F. S. Natush, John R. Wallace and Charles A.
Evans, Jr., "Trace Elements in Flue Ash, Dependence of Concentration on
Particle Size," Env. Sci. Tech., 8 (13), pp. 1107-1113, December 1974.
5. Telephone conversation with M. H. Schwartz - Chemical Engineering Depart-
ment -Shell Development Company, Houston, Texas, May 16, 1977.
6. Cuffe, S. T., et al., "Air Pollutants Emissions from Coal-Fired Power
Plants," Reports 1 & 2, 56th Annual Meeting APCA, Detroit, Michigan, 1963.
7. Hollinden, G. A., et al., "Control of NOX Formation in Wall Coal-Fired
Boilers," Preceedings of the Stationary Source Combustion Symposium
Vol. II, EPA-600/2-76-152b, June 1976, Atlanta, Georgia.
8. Cowherd, C., et al., "Hazardous Emission Characterization of Utility
Boilers," EPA-650/2-75-066, July 1975.
9. Crawford, A. R., Manny, E. H., Bartok, W., Exxon Draft Report - To be
released.
10. Grouhoud, et al., "Some Studies on Stack Emissions from Lignite-Fired
Power Plants," Technology and Use of Lignite, Bu-Mines-IC 8650, Oct. 1, 1974.
11. Mesich, F. G., et al., "Coal-Fired Power Plant Trace Element Study,
Station 1, 2 & 3," Radian Corporation, EPA Contract No. 68-01-2663, 1975.
12. Selker, A. P., "Program for Reduction of N0x,from Tangential Coal-Fired
Boilers -Phase II," EPA-650/2-73-005, June 1975.
13. U.S. Federal Register Vol. 36, No. 247, December 23, 1971.
14. Burnington, R. L., et al., "Field Test Program to Study Staged Combustion
Technology for Tangentially Fired Utility Boilers Burning Western U S
Coal Types," Draft Report, Combustion Engineering, Inc.
65
-------
15. Ctvrtnicek, T. E., et al., "Evaluation of Coal-Sulfur Western Coal Char-
acterization, Utilization, and Distribution Experience," Monsanto
Research Corp. EPA 650/2-75-046, May 1975.
16. Crawford, A., et al., "Field Testing: Application of Combustion Modifi-
cations to Control NO Emission from Utility Boilers," EPA-650/2-74-066,
June 1974. x
17. Gregory, M. W., et al., "Determination of the Magnitude of S02, NO, C02,
and 0£ Stratifications in the Ducting of Fossil Fuel-Fired Power
Plants," Exxon Research and Engineering Co., Presented at the 69th Annual
Conference of the APCA, June 27-July 1, 1976.
18. Personal Communication with Frank Sustino, Combustion Engineering Inc.,
September 23, 1977.
19. Beck, A. A., and Burdick, "A Method of Test for S02 and SOs in Flue
Gases," Bureau of Mines Report of Investigations 4818, January 1950.
20. Smith, W. S., and Gruber, C., "Atmospheric Emissions from Coal Combustion -
An Inventory Guide," PHS Rep. 999-AP-24.
21. Smith, J. F., "Sampling and Analytical Modifications of the Beck and
Burdick Method For SOe and SOs Analysis," Bureau of Mines.
22. Goks0yz, H., and Ross, K., "The Relation Between Acid Dew Point and
Sulfur Trioxide Content of Combustion Gases," Thornton Research Center,
Shell Research, Ltd., 1962.
23. Lisle, E.S., and Sensenbaugh, J. D., "The Determination of Sulfur
Trioxide and Acid Dew Point in Flue Gases," Combustion Engineering, Inc.,
Combustion, January 1965.
24. Wohlschlegel, P., "Guidelines for Development of A Quality Assurance Pro-
gram: Volume XV - Determination of Sulfur Dioxide Emissions From
Stationary Sources By Continuous Monitors," EPA-650/4-74-005-0, March
1976.
66
-------
APPENDIX A
MATHEMATICAL RELATIONSHIPS USED
1. 100 percent fuel sulfur conversion to S02
S02 (^) = 8.598 x 106 ^
S = percent sulfur in the coal
HV = heating value of the coal as fired (Btu/lb)
2. Conversion of sulfur dioxide emissions from Ib/HR to ng/J
S02 (*f) - 4.299 X !0« (S02 %) (HV) fa)
HV = heating value of the coal as fired (Btu/lb)
FF = coal flow (Ib/hr) ;
3. Conversion of ppm S02 to ng/J
S02 (^ = 4.299 x 102 (MsOo) (ppm S02)
*•
HV
Mso2 = molecular weight of S02 = 64
HV = heating value of the coal (Btu/lb)
= moles of dry flue gas per pound of fuel (dry basis)
_ 4.762 (nc + ns) + 0.9405 nH - 3.762 no2 fuel
1 - 4.762 %02
100
% carbon in the coal (as fired)
nc " 1200
_ % sulfur in the coal (as fired)
nS - 3200
nf
9d
67
-------
S02
S02
_ % hydrogen in coal (as fired)
nH ~ - 100
_ % 02 in the coal (as fired)
2 '3200
% 0? = percentage excess oxygen in the stack
= 2.751 x 101* (ppm S02) I^
4. SO
= 3.439 x 10* (ppm S03)
= 4.299 x 10* 1b N°x
S0
S0
MBtu ppm NO MW NO
MW S02 = molecular weight of S02 = 64
MW NO = molecular weight of NO = 30
5. Percent fuel sulfur from ash free basis to total weight percent basis:
/T r»rt o/-» « U \
%S = %S(ash free basis)
\ • w /
Ib SOx
QQ
6. S0x Hi = 43° ^ 1Q3 [l + 206 (nc + ns) + 42/5 nH 1
A J HV |_ J
1000 Ib dry flue gas 50%/EA
68
-------
APPENDIX B
COMPARISON OF S02 EMISSION FACTORS
Table B-l presents the emission factors listed in U.S. EPA AP-42
together with the emission factors obtained in this study. The two sets
of data compare favorably except for the emission factors for the high
sodium ash lignite fired boilers. The emission factor reported in this
study for lignite represents an average of all the readily available data
from high sodium lignitic coal. If only the data from the Hoot Lake boiler
(Figure 6) are considered, then the conversion becomes approximately 50
percent. The resulting emission factor of 20 S compares more favorably with
the EPA value. It is believed that the Hoot Lake data might be more reliable
than the overall average, since the tests were conducted specifically to
measure the effect of sodium in the ash on S02 emissions.
69
-------
CO
g
CO
CO
LU
UJ
Q
t—I
X
o
o;
g
co
CO
CD
LU
_J
CO
cu
o
E E
CU O
S- -l->
4— -Q^
q-r-
Q
g
c-
4J
O
S_
cu
«=C
CM
^r
l
O_
"^
to
1 % f»
E O
CUv-
r— v> o
CU 4->
•i- > fS
3 E "^
cr o
ui to
E
0
i^_
o
g__
i— 0)
>
co
CO
o
S;
CO
CO
CO
co
If)
_Q
co
CO
CO
co
3
o
•l_
pz
3
•r—
CO
CO
«=C O '
CM
co
co co
CO ID
co co
3
O
r—
CU . -E
4-> 1 IO
•r- fO
O CU
(d 4^> E
S- -1-3
.E E •!-
4-5 01T3
E -i- O
«C . — 1 W
C\J
V
CU
T3
X
0
E
•I—
-o
O
in
CO
o
,_!
+
CO
CO
, —
CO
LO
|*^»
co
CD
co
gjZ
3
•r- CU
-0 4-»
. O T-
t/> E
Ol
E-r-
•r-i
•a x:
CU CO
^^ rd
CO C/)
0 0
en "=l-
+ i
10 CO
CO CO
10 •*
CM CO
in
•
CM un
«^ cr*
CO CO
r^ co
t— CO
-E E
10 3
cd -i- «o
-a 3
E O 0
3 CO- — • E
•r— ••— ^^^ *r~
T3 CO E
O CU A 3
10 4-> 4->
•r- 0) -r-
^: E -a -Q
Co cn*r~ *~*
•r- -i- X 3
3= c— 0 CO
CO
cu
3
(d
cu
en
(d
s-
cu
rd
4->
^
CU
to
cu
S-
a.
cu
i^
fd
^9
(d
^"^i
-------
APPENDIX C
INSTRUMENTATION AND SAMPLING TECHNIQUES
Table C-l lists the instrumentation and sampling techniques used to
insure gaseous sulfur emissions data reported in Section 3. The equipment
varied significantly among the test programs, thereby introducing another
variable when comparing sulfur oxides data.
The methods used can be divided into two main groups:
a Wet chemistry (grab sample)
« Electronic monitors (continuous sample and intermittent grab
sample)
The wet chemistry methods include the EPA Reference Method No. 6, the Berk
and Burdick and the Lisle and Sensenbough methods. The continuous monitor-
ing techniques include the extractive ultraviolet absorption and the non-
dispersive infrared (NDIR) methods.
All these methods are comparable in measurement accuracy; however,
they all require different sampling procedures, which can be the source of
possible errors if appropriate precautions are not taken.
For example, the wet chemical methods involve the use of sampling
trains which grab a predetermined flue gas sample for chemical analysis,
usually by titration method. The grab sample is most often taken from a
single location in the stack, usually 2 to 3 feet from stack walls. This
single point sample can be nonrepresentative of the average sulfur oxide
concentration due to gaseous stratification. Typical errors caused by single
point sampling are ±20 percent but can be as high as ±48 percent of the
measured value (Reference 17). Sulfur oxide data from single point sampling
were reported in References 6, 10, 12 and 14.
71
-------
S9S-1
a.
to
i
to
eC
CO
on
o
to
rD
oo
UJ
s-
I—I
z=
3=
I
UJ
to
UJ
a!
1
fi
tn
gQJ
t/i
11
||
£^
4-* cr
MUJ
c
h^
CI
i
t/t
o
i
r— 3
H
u
!l
if
to
Modification wade to eliminate
Interference on SO? readings.
Hydrochloric add titration for
SO? analysis and standard ben-
zene method for SOa analysis.
1
u
11
XJ
E
CO
CM
o
t/]
_£
U COO)
t- 4J 01 IM «
*J C cnt— o
1 1 I 1
a^s s =
S x a a
IIII
^
J=
re
If. (U
O o
tfl U (U
+j- 3 x:
OO-O 4->
«3 OJ •*—
cn-
3 tn
tf. QJ Q}
4-> X) U
5 3 2 O.
J—T3 O- tO
4->
C
ICO
-M -M a
0 W S
"•" """o
O 4-» C C3
nj ax^r
*3'o CM •
x •»- o o
Ul >
I
•M
^H
c?1
to
tn
o
2=
«1
O X.
il
CO
S02 was collected 1n Impinger
I No. 3 which contained 10%
sodium carbonate.
c
T- cn
(/i «c e
3 O.f-
LU (—
* O.
to -o E
QJ re
-a^- to
x: T-
&1IS
XI
£
0~
tn
o
2=
»
re t *o.
re x: T3 o
3 4-> C C 0) 3
T- tn
• -o x to to
Cn QJ
s- s_ tn c
3 a>-r- co
tS-'S
3 OJ >j QJ
(o -a x» to
1
C3
o
VI
CM
3 QJ*O re
I"!1 1
Y°>,|i
(Ur— C • Q)
5C O O O£ J^
N as
p
For Stations No. 1 and No, 2
samples were collected from the
stack downstream of the scrubber
(Station No. 1) or ESP (Station
No. 2). For Station No. 3 sample
was collected from stack leaving
cyclone collectors.
QJ
•o o
to t. o>
OJ tn *
•T) 0. 3 CM
O O
JC C 10 tO
+J Q) £_
uj ic xi re
5-
CO
t/t
OJ
|
1 1 !
r— CM CO
o re o re o
= 5 = 5==
l/» C C
§QJ QJ
cn cn
-p r- U> i—
re o *
E OJ QJ O O
£OJ i— > 2=
r— O.I— O
H- CT> E 0> 4J >>
c re 3 S-
-O -r- (/) J_ -O t-
O) tO OJ QJ
4-> ^O tn CQ
o re QJ • 3
re z: (/>
Q) +J t/) QJ
r— ai cn >>4-> •—
§•££!:£§
*a 3 w a) o re
to o 3 ca a. tn
' -
£-
1 i.
£= QJ
ro cm
to re
I_ E t.
O CM 2 O
4- 0<4-
* QJ
to o OQ -a
z: aire
-a s- •
o >j • re «*•
x: s- to s-
OJ OJ t— C O
E CQ OUT- SC
CM >>
• IT
£m
£o
(U
CO V)
s- o
O 3
*»- C
re e
s- o
CD O
CM
O
to
re re
e c
•U QJ
0, g,
re re
i— i—
i t
CM ^3-
O O
& &
(U (U
CO CO
«a-
£^
Q) C=
to
S- 4->
r^ QJ
O.4-> C
£ OJ T-
to •(-»
o QJ re
Q. N 10
*cn c -*J
c o re
•i- O OJ
to oj x:
^
Ul
'E
?!
X!
U
•M
£
to
a~
.
J. §3
Columbia No.
Tangential
Huntlngton Ci
2 - Tangen
cu
u
I
-------
Multiple point sampling using EPA Method No. 6 was used during test
programs reported in References 8, 9, and 11.*
In these test programs, a combined EPA Method 5 (particulate test)
and Method 6 were combined by changing the situations in the impingers from
distilled water to hydrogen peroxide and isopropyl alcohol (as described in
EPA Reference Method 8 - Reference 13).
Another source of error associated with grab sampling comes from
sample handling and analysis. Errors due to these operations can be very
significant if contamination is not avoided and prescribed sample proce-
dures are not followed closely. Unfortunately these errors are impossible
to identify and quantify because fully documented procedures for each of
these test programs are not available.
Continuous monitors were used to collect S0? data from only two
sources, namely Barry No. 4 (Reference 12) and Willows Creek No. 5 (Reference
7). In the case of Barry No. 4, the use of continuous monitors permitted
the measurement of sulfur dioxide from a composite of 12 individual flue gas
samples. In the case of the Willows Creek No. 5 tests, it is believed that
intermittent grab samples were taken from six individual test points. This
assumption of continuous grab samples is based on the fact that the ultra-
violet adsorption instrument analyzes one grab sample at a time.
One common source of error for these two analyzers, as with all
electronic analyzers in general, is in the calibration of the instrument.'
Proper calibration procedures are necessary to account for changes in
instrument response caused by drift, instrument wear and analyzer contami-
nation. Another source of error associated with the NDIR alone is in the
sample handling and conditioning interface necessary with the use of this
instrument. The interface removes particulate and moisture from the flue
gas sample prior to exposing the sample to the sensor. This interface can
be a source of errors because of leaks or doesn't provide sufficient con-
ditioning.
*
EPA Method 6 does not specifically require traversing the stack. However,
composite samples might have been taken because Methods 5 and 6 were com-
bined to measure particulate and sulfur emissions during these test programs.
73
-------
The UV analyzer is usually located next to the stack; therefore, it
avoids the use of long, potentially leaky sample lines. Furthermore, this
instrument does not require the removal of moisture from the flue gas as
long as the sample is maintained above its water dew point.
In conclusion, measuring error probably caused many of the data to
show conversion of over 100 percent. The most easily identified error
is that due to single point grab sampling instead of multiple point sampling
or traversing. Other errors stemming from instrument operation, sample
handling, and fuel sampling and analysis are difficult to identify, so
they can only be speculated upon.
A brief description of the type of instrumentation used in each of
the field tests investigated follows.
Modified Berk and Burdick Method
The Berk and Burdick Method used in Reference 6 uses an acidimeter
type of analysis for determining S02 and SOs emissions in power plant
effluents. The original method described in Reference 19 was shown to
have interferences in the analysis of S02 when acid gases such as N02, HCE,
NHs and organic acid were also present in the measured gases.* This inter-
ference caused the S02 readings to be 15 to 50 percent higher than the
theoretically expected values.
However, the reported S02» $03 emissions from Units "A" through "D"
in Reference 6 were measured using a modified version of the Berk and
Burdick Method. The modification consisted of using hydrochloric and
benzidine solutions when titrating the flue gas samples. These solutions
eliminate the interference of CL, 1% and organic acid gases (Reference 21).
Lisle and Sensenbough Method
For the tests performed on the Hoot Lake, Milton N. Young, Lei and
Olds and William J. Neal boilers (Reference 10), a modified sulfur oxide
condenser was used. This condenser was first devised by Goksoyn and Ross
(Reference 22) and later investigated by Lisle and Sensenbough (Reference 23).
*
These acids are quite frequently found in flue gases from combustion of
coal (Reference 20).
74
-------
The intake apparatus consists of a glass heated probe followed by a helical
glass coil and a glass fit. The coil is immersed in a heated water jacket
which permits moisturizing the condenser temperature between the acid dew
point and the water dew point. Since the acid dew point can be defined as
"the temperature at which the combustion gases are saturated with sulfuric
acid," then the dew point-acid concentration relationship can be determined
for known amounts of sulfur oxide inlets to the condenser. This relation-
ship is then used to determine unknown concentrations of sulfur oxide based
on the flue gas dew points. The demonstrated accuracy of this apparatus
for SO., measurement has been reported to be ±0.3 ppm in the range of concen-
trations normally encountered in stack flue gases.
Ultraviolet Absorption Method
The extractive ultraviolet absorption method employed in Reference 7
consists of measuring electrical signals-generated by wavelength photo-
tubes which measure intensity of light beams. The instrument uses a
sample and a reference light beam. Sample gases containing S02 are passed
through the sample beam. S02 absorbs light at certain wavelength causing
a change in intensity of the beam. The change in intensity is detected by
the phototube which in turn releases an electrical signal proportional to
the concentration of the S02 in the gas.
Nondispersive Infrared Method
The nondispersive infrared analyzer (NDIR) used to measure SO,,
emissions from Barry No. 2 (Reference 12) is the most common continuous
monitoring technique for S02 measurement.
The NDIR technique consists of either one light source with a light
chopper or two identical sources whose beams are directed through two
different cells. One of the cells contains a gas which does not absorb
infrared energy at the same wavelengths at which sulfur dioxide absorbs
infrared energy. Passing through the other cell is the sampled stack
gas. The beams pass through both of these cells and into different half
sections of a reference chamber. Separating the two half sections of the .
reference chamber is a flexible metal diaphragm. Both sections contain the
75
-------
same amount of S02 vapor kept at the same atmospheric pressure. The degree
of absorption of infrared energy by the sample gas is directly proportional
to the amount of S02 in the sample gas. The absorption by the sample gas
mil proportionally reduce the absorption by the S02 vapor in the corres-
ponding half section of the reference chamber. The difference between the
energy absorptions in the two halves on the reference chamber, then, is a
measurement of the concentration of S02 in the sample gas.
The primary sources of error in the NDIR method are the blocking of
the transmission of the light beam by particulates and the inadvertent
absorption of infrared energy by moisture in the sample gas. Both of these
sources of error can be minimized by adequate inferfacing equipment.
The sampling interface used with an NDIR analyzer must be capable of
removing flyash and particulate matter as well as removing or determining
the quantity of moisture in the sample. Particulate matter will tend to
collect on the windows of the sample cell. Water vapor will interfere
inasmuch as the SCL absorption band is overlapped by a water system in the
1200-cm"1 to 1400-cnr1 region (Reference 24).
U.S. EPA Method 6
This method uses a glass probe followed by a set of four impingers
immersed in an ice bath. A gas sample is extracted from the sampling point
in the stack. The sulfuric acid mist, including S03 and S02, are separated.
S03 is collected in the first impinger bubbles containing isopropyl alcohol
solution, while SCL is collected in the following two impingers containing
an hydrogen peroxide solution. Possible interference due to cations and
fluorides in the flue gas are eliminated by inserting a glass wool filter
in the probe. The probe is maintained at a temperature higher than the dew
point of the water in the flue gas. The samples are titrated with the
Barium-thorin method to measure S02 and S03-
76
-------
APPENDIX D
TABLE OF CONVERSION UNITS
SI Metric to English Conversion Factors
To Convert From
J/g
MJ/S
ng/J
kg/S
To
Btu/lb
TO6 Btu/hr
lb/106 Btu
TO3 Ibs/hr
Multiply by
4.299 x 10'1
3.412
2.326 x TO'3
7.936
English to SI Metric Conversion Factors
To Convert From To Multiply by
Btu J 1.0548 x 103
Btu/lb J/g 2.326
106 Btu/hr MJ/S 2.9307 x 10'1
lb/106 Btu ng/J 4.299 x 102
103 Ibs/hr kg/S 1.26 x TO"1
MW (electrical) J/hr 1.0548xl010
(assumes 34 percent- plant
efficiency)
J = Joule
g = gram
S = second
W = watts
M = mega (106) .
n = nano (10~9)
k = kilo (103)
77
-------
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
2.
3. RECIPIENT'S ACCESSION'NO.
4. TITLE AND SUBTITLE
Boiler Design and Operating Variables Affecting
Uncontrolled Sulfur Emissions from Pulverized
Coal-Fired Steam Generators
5. REPORT DATE
February. 1978
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Carlo Castaldini & Meredith Angwin
. PERFORMING ORGANIZATION NAME AND ADDRESS
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, California 94042
1O. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2611
2. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Office of Air Quality Planning .and Standards (MD-13)
Research Triangle Park,. North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final '
14. SPONSORING AGENCY CODE
5. SUPPLEMENTARY NOTES
6. ABSTRACT
The report presents an analysis of the data from eight field test reports for
twenty-one steam generator/coal type combinations.^ The data were analyzed to
determine boiler design and operating variables which affect SQ? emissions, the
extent to which emissions were affected, and trends in conversion of sulfur in coal
to S02, S03, and solid sulfates.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
13. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)'
Unclassified
21. NO. OF PAGES
80
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
79
-------
-------