-------
80
60
<
o
o
•H
i-,
20
0
1969 1971 1973 1975
Figure 7 Price of by-products
QuantityC millions of tons)
o ru •£• <^
-
Demand
OU
B
C
Supply
OS
R
P
Demand
OU
B
C
>>
,-H
O<
DH
3
to
OS
R
P
Demand
OU
B
C
Supply
OS
R
P
1970
1973
1975
Demand: C:Cement B:Board OU:Other uses
Supply: P:Phosphogypsum R:Recovered OS:Other sources
Figure 8 Demand for and supply of gypsum in Japan
75
-------
The by-product gypsum from flue gas from coal-fired boilers at Omuta
plant, Mitsui Aluminum and Takasago plant, E.P.D.C. contains less
than about 6$ impurities (fly ash, limestone, etc.) and has been
used mainly for cement and partially for wallboard. Isogo plant,
E.P.D.C. and Wakamatsu plant, Nippon Steel, both under construction,
will produce gypsum with more than 20$ impurities to be discarded.
The first unit of Mitsui Aluminum in operation since 1972 by-producing
calcium sulfite sludge will have to be modified to by-produce gypsum
because the sludge pond is nearly full. Although gypsum is in over-
supply, throw-away calcium sulfate sludge will not increase because
the sludge needs much more land space than does gypsum.
Elemental sulfur is a desirable by-product but FGD processes by-
producing sulfur seem very expensive except for oil refineries that
have Glaus furnaces. To by-produce elemental sulfur, hydrodesulfuri-
zation of heavy oil may be more economical while it is more expensive
than FGD by the wet lime/limestone process (Table 9).
Table 9 Cost comparison at various desulfurization ratios
(f/kl, at 6,000 hours a.*»mal operation)
FGD (wet-lime/limestone process)
Hydrodesulfurization of oil
70*
17
17
809&
18
19
90%
19
22
97$
20
27
Two commercial hydrodesulfurization plants recently started operation
to reduce sulfur to below 0.3$ (over 90$ removal) using heavy oil
with little metallic impurities which tend to poison the catalyst.
Most heavy oils are rich in impurities and difficult to desulfurize
over 80$. For such heavy oils, vacuum distillation followed by
hydrodesulfurization of the distillate and decomposition of the re-
sidue with partial gasification has been considered rational. The
first commercial plant with the decomposition started operation in
January 1976 using the Xurena process (or Eureka process). The
second plant is near completion using the Flexicoking process.
76
-------
5.2 Simultaneous removal
Fuji Kasui claims that the plant cost for simultaneous removal ranges
from $60 to 90/kW and the operation cost is roughly f 30/kl oil in-
cluding depreciation for 7 years. The costs for other processes are
uncertain because those are still in a test stage.
Activated carbon processes have an advantage in that they are a dry
process and can be operated at relatively low temperatures. Carbon
for simultaneous removal is fairly expensive at present-, viz., about
$8,000/t, while carbon for PGD costs about $3,000/t. To treat flue
gas from a 350MW boiler at a SY of 1,000, about 1,000 tons of carbon is
required. Production of much cheaper activated carbon is desired.
The Shell process seems costly for PGD but its capability of simul-
taneous removal may compensate for the disadvantage.
Among the by-products from wet processes, sodium nitrate and nitrite
have little use. Calcium nitrate may be used as fertilizer but its
demand is limited due to the low nitrogen content and high hygroeco-
picity. Ammonium nitrate and nitric acid are better by-products.
By the hydrodesulfurization of heavy oil, 30-40$ of nitrogen in the
oil is also removed resulting in 10-20$ abatement of NOX in flue gas.
Tests have been made to remove much more nitrogen in the oil along
with sulfur. To compete with possible low-sulfur low-nitrogen oil,
the operation cost of simultaneous removal from flue gas should be
less than about $30/kl and the process must not have difficulty in
wastewater treatment and disposal of by-products.
Efforts for denitrification in Japan has been exerted mainly for flue
gas from oil burning containing 200-300ppm NOX because oil is the
major fuel in the country. The need for denitrification may be larger
for flue gas from coal burning which normally contains more than
600ppm NO which is not easy to reduce to below 400ppm by combustion
modifications. Denitrification will become more and more important
in many countries as the consumption of fossil fuels increases.
77
-------
Conversion figures
1 m^ = 35.3 cubic feet
1 liter =0.26 gallon
1 kl = 6.29 barrels
1 NnrVhr = 0.^9 scfm
L/G 1 liter/Nm^ =7.4 gallon/1,OOOscf
References
l) J. Ando, Recent Developments in Desulfurization in Japan (January 1975)
Being published by U.S. EPA through PEDCO.
2) G. A. Hollinden and F. T. Princiotta, Sulfur Oxide Control
Technology, Visits in Japan (March 1974)* Interagency Technical
Committee.
3) J. Ando and H. Tohata, HOx Abatement Technology in Japan for
Stationary Sources (March 1975). Being published by U.S. EPA
through PEDCO.
4) M. Seki, Y. Sakurai and K. Yoshida, Ammonium Halide-Activated
Carbon Catalyst to Decompose NOx in Stack Gases at Low Temperatures
around 100°C. American Chemical Society (August 1975. at Chicago).
5) S. Yamada, Y. Watanabe and H. Uchiyama, Bench-Scale Tests on
Simultaneous Removal of SO* and NOX by wet Lime and Gypsum Process
Ishikawajima-Harima Engineering Review, Jan. 197&, Vol. 16, No.i.
(In Japanese)
78
-------
FLUE GAS DESULFURIZATION ECONOMICS
G. G. McGlamery, H. L. Faucett, R. L. Torstrick, and L. J. Henson
Office of Agricultural and Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
ABSTRACT
Described herein are several phases of flue gas desulfurization
(FGD) process and byproduct economic evaluations being carried out
by the TVA Office of Agricultural and Chemical Development for EPA
and the TVA Office of Power. Included are recently updated investment
(mid-1977) and annual revenue requirements (mid-1978) for five leading
FGD processes evaluated by TVA in 1974. Also included are results
from a cost sensitivity study of limestone scrubbing options based
on data from the EPA-Bechtel Corporation-TVA test program at Shawnee,
plus some preliminary evaluations of promising double-alkali and
regenerable processes.
In addition to these results, several new or expanded evaluation
programs are described. Included are:
1. A computer program prepared by Bechtel and TVA to screen process
options of lime-limestone scrubbing using Shawnee data.
2. Marketing studies covering volume, pricing, and location aspects
of various FGD byproducts.
3. Evaluations of advanced FGD processes, N0x control processes, and
systems for clean fuel from coal.
4. A sludge disposal system design and cost study.
5. Process energy optimization studies.
The status of these programs is also discussed; some phases of
these programs are nearing completion, while others are either just
getting under way or yet to be started.
79
-------
FLUE GAS DESULFURIZATION ECONOMICS
INTRODUCTION
Throughout the developmental stages of flue gas desulfurization
(FGD) technology, process economics has been one of the more interesting and
debatable subjects, to say the least. From the earliest cost estimates of
the late 1960's to now, and no doubt in the future, the estimated and actual
costs of such systems have been increasing to higher and higher levels. But,
so has the cost of most everything else in today's inflation-ridden society.
Nevertheless, not all of the increases in cost estimates for FGD systems can
be attributed to inflation, since like most new emerging technologies, as
experience and knowledge increase in the early years of application, so does
the expected real costs of such facilities.
Some perspective can be gained from Figures 1 and 2 which show the
capital cost of the generalized limestone slurry scrubbing process as esti-
mated by TVA over the past decade. Although the scope of these estimates has
tended to change slightly during this period, the increases are due to both
improvements in knowledge of process requirements and inflated labor and
equipment costs.
By assuming a relatively simple example case, we have been fairly
successful in assessing the comparative economics of FGD systems. However,
on a total magnitude basis, it has been much more difficult to predict repre-
sentative costs of specific systems. For that matter, estimators throughout
the industry have encountered this difficulty, primarily for the following
reasons.
1. Variable project scope, time frame, design, and operating conditions.
2. Numerous process options.
3. Vendor optimism and user skepticism.
4. Confusion over regulatory requirements.
5. Demanding physical and chemical environment.
6. Variable byproducts and byproduct value/disposal cost.
7. Shortage of actual design and operating data.
At least for the near future-, we expect most of these reasons to remain as
impediments to total cost assessment.
80
-------
too
80
T
T
500 MW, 3.5% S coal-fired units
Generalized TVA estimates
CO
01
(U
c
60
3 40
20
Scope and technology
1968
70
72
74
Year
76
78
Figure 1. Limestone slurry scrubbing process -
change in investment cost during past decade.
80
-------
00
N)
100
0)
80
60
i
I 40
20
3.5% S coal-fired units
Generalized TVA estimates
Mid-1977
Mid-1972
1
1
1
200 400 6OO 800
Power unit size, MW
IOOO
Figure 2. Limestone slurry scrubbing process -
range of projected investment estimates.
1200
-------
UPDATE OF THE PAPER "COST COMPARISONS OF FLUE GAS DESULFURIZATION SYSTEMS"
Since 1967 the TVA Office of Agricultural and Chemical Development
has been preparing for EPA or its predecessors cost estimates of various FGD
processes. Many of these have been published or presented previously. Prob-
ably the best known set of EPA-TVA FGD process cost estimates was presented
ia a paper entitled "Cost Comparisons of Flue Gas Desulfurization Systems"
at the EPA symposium in Atlanta, November 1974. Five processes were evaluated
in a single comparative study—limestone slurry and lime slurry scrubbing;
•agnesia slurry - regeneration to 98% sulfuric acid; sodium scrubbing -
regeneration to sulfur; and catalytic oxidation to 80% sulfuric acid. The
complete data from which the paper was prepared were later published in
January 1975 in an EPA report entitled "Detailed Cost Estimates for Advanced
Effluent Desulfurization Processes." Because the project time frame for these
estimates was mid-1972 to mid-1975 and capital and revenue requirements were
subjected to tremendous escalation near the end of this period, TVA has up-
dated these costs to reflect a mid-1975 to mid-1978 construction project.
Ihe revised investment costs are shown in Table 1 and the annual revenue
requirements are shown in Table 2. (Revenue requirements include all oper-
ating costs, such as raw materials, labor, utilities, maintenance, overheads,
capital charges, and taxes; assumes startup in mid-1978.) The basis and data
used in updating these values are given in Table 3.
It should be noted that the revenue requirements shown in Table 2
do not reflect byproduct credits or sludge fixation costs. With recent in-
creases in natural gas prices affecting the Frasch sulfur mining costs, we
have seen sulfur list prices jump 85% (from $35 to $65/long ton) in 1975 with
subsequent effect on sulfuric acid list prices (from $30 to $45/ton, 100%
acid). Although these markets are now temporarily soft and spot prices re-
flect discounting, on a long-term basis, elemental sulfur prices can be
expected to continue to climb with acid prices to follow. Introduction of
large quantities of abatement byproducts would certainly change the picture,
tat if smaller quantities are introduced over a reasonable period of time
(at the same rate as the total market growth), market disruption could be
held to a minimum. Based on recent work (paper entitled "Potential Utiliza-
tion of Controlled SOX Emissions from Power Plants in Eastern United States"
by J. I. Bucy, et al., presented at this meeting) TVA has completed for EPA
on byproduct marketing, it would not be unreasonable to expect a long-term
net revenue of $45/long ton of sulfur or $25/ton of 100% sulfuric acid.
Recent EPA policy statements now seem to be encouraging sludge
fixation or similar long-term sludge disposal practices. If such costs must
be added to those projected in Table 2 for limestone and lime scrubbing, the
co^>etitive position of these processes with respect to regenerative processes
could change drastically. As illustrated in Table 4, the effect of byproduct
credits for the regenerable processes and sludge fixation costs for the non-
regenerable systems are significant. Since 90% or more of the FGD systems
operating or under construction are the nonregenerative type, the utilities
obviously need to take a closer look at the regenerative systems.
83
-------
Table 1. SUMMARY OF TOTAL CAPITAL INVESTMENT REQUIREMENTS
FIVE LEADING FLUE GAS DESULFURIZATION PROCESSES
1975-78 COST BASIS
a,b,c
00
Case
Coal-fired power unit
90% S02 removal; onsite solids disposal6
200 MW, new, 3.5% S
200 MW, existing, 3.5 %S
500 MW, existing, 3.5% S
500 MW, new, 2.0% S
500 MW, new, 3.5% S
500 MW, new, 5.0% S
1,000 MW, existing, 3.5% S
1,000 MW, new, 3.57, S
80% S02 removal; onsite solids disposal6
500 MW, new, 3.5% S
90% S02 removal; onsite solids disposal6
(existing unit without existing
particulate collection facilities)
500 MW, existing, 3.5% S
Oil-fired power unit
90% S02 removal; onsite solids disposal6
200 MW, new, 2.5% S
500 MW, new, 1.0% S
500 MW, new, 2.5% S
500 MW, new, 4.0% S
500 MW, existing, 2.5% S
1,000 MW, new, 2.5% S
Years
life
30
20
25
30
30
30
25
30
30
25
30
30
30
30
25
30
Limestone
M $
17,671
15,282
31,113
30,734
34,185
37,114
47,529
51,362
32,955
40,314
11,278
17,778
21,189
23,889
25,215
32,866
$/kW
88.4
76.4
62.2
61.5
68.4
74.2
47.5
51.4
65.9
80.6
56.4
35.6
42.4
47.8
50.4
32.9
Lime
M $
15,978
17,527
35,001
27,545
30,558
33,143
51,620
44,891
29,428
35,119
12,892
21,739
24,778
27,149
29,329
36,193
$/kW
79.9
87.6
70.0
55.1
61.1
66.3
51.6
44.9
58.9
70.2
64.5
43.5
49.6
54.3
58.7
36.2
Magnesia
M $
19,119
19,210
34,940
31,057
35,827
39,941
52,293
53,013
34,659
43,145
12,002
17,224
21,871
25,588
27,309
32,377
$/kW
95.6
96.1
69.9
62.1
71.7
79.9
5£.3
53.0
69.3
86.3
60.0
34.4
43.7
51.2
54.6
32.4
Sodium
M $
22,391
23,597
43,351
36,802
42,387
47,210
67,038
64,203
40,406
52,419
14,335
21,001
26,500
30,925
33,618
40,590
$/kW
112.0
118.0
86.7
73.6
84.8
94.4
67.0
64.2
80.8
104.8
71.7
42.0
53.0
61.9
67.2
40.6
Cat -Ox
M $
25,171
22,914
48,772
54,494
54,791
55,047
80,497
88,893
-
56,585
16,665
35,552
35,814
36,014
42,366
58,122
$/kW
125.9
114.6 .
97.5
109.0
109.6
110.1
80.5
88.9
-
113.2
83.3
71.1
71.6
72.0
84.7
58.1
Midwest plant location represents project beginning mid-1975, ending mid-1978. Average cost basis for scaling, mid-
1977. Minimum in process storage; only pumps are spared. Investment requirements for disposal of flyash excluded.
Construction labor shortages with accompanying overtime pay incentive not considered.
These investment costs depend heavily on project definition.
^ Working capital has been included in the total capital investment.
All Cat-Ox installations require particulate removal to 0.005 gr/acf prior to entering converter. Because existing
units are assumed to already meet EPA standards (0.1 Ib particulate/MM Btu of heat input), only incremental additional
precipitator is required.
Sludge disposal pond with clay liner.
-------
00
tn
Table 2. SUMMARY OF TOTAL AVERAGE ANNUAL REVENUE REQUIREMENTS
FIVE LEADING FLUE GAS DESULFURIZATION PROCESSES
1978 COST BASIS
Total average annual revenue requirements
a,b
Case
Coal-fired power unit
90% S02 removal; onsite solids disposal
200 MW, new, 3.5% S
200 MW, existing, 3.5% S
500 MW, existing, 3.5% S
500 MW, new, 2.0% S
500 MW, new, 3.5% S
500 MW, new, 5.0% S
1,000 MW, existing, 3. 5% S
1,000 MW, new, 3.5% S
80% S02 removal; onsite solids disposal
500 MW, new, 3.5% S
90% S02 removal; onsite solids disposal
(existing unit without existing
particulate collection facilities)
500 MW, existing, 3.5% S
Oil-fired power unit
90% S02 removal; onsite solids disposal
200 MW, new, 2.5% S
500 MW, new, 1.0% S
500 MW, new, 2.5% S
500 MW, new, 4.0% S
500 MW, existing, 2.5% S
1,000 MW, new, 2.5% S
Years
life
30
20
25
30
30
30
25
30
30
25
30
30
30
30
25
30
Limestone
M $
5,883
5,686
11,854
10,625
11,937
13,105
19,711
19,163
11,319
14,376
4,236
7,377
8,548
9,566
9,979
13,317
Mills/kWh
4.20
4.06
3.39
3.04
3.41
3.74
2.82
2.74
3.23
4.11
3.02
2.11
2.44
2.73
2.85
1.90
M $
6,362
7,150
14,528
11,145
12,758
14,377
23,819
20,570
12,206
14,826
5,123
8,920
10,585
11,886
12,140
17,165
Lime
Mills/kWh
4.54
5. 11
4.15
3.18
3.65
4.11
3.40
2.94
3.49
4.24
3.66
2.55
3.02
3.40
3.47
2.45
Magnesia
M $
7,036
7,257
14,052
11,572
14,082
16,448
23,169
22,789
13,406
16,639
4,615
6,848
9,019
10,967
10,683
14,734
Mllls/kWh
5.03
5.18
4.01
3.31
4.02
4.70
3.31
3.26
3.83
4.75
3.30
1.96
2.58
3.13
3.05
2.10
Sodium
M $
9,238
10,868
22,189
14,549
18,782
22,858
38,813
31, 186
17,425
24,995
6,463
8,938
1.3,075
17,042
15,322
22,223
Mills/kWh
6.60
7.76
6.34
4.16
5.37
6.53
5.54
4.46
4.98
7.14
4.62
2.55
3.74
4.87
4.38
3.17
Cat-Ox
M $
6,000
8,263
17,765
12,681
12,766
12,844
31,138
20,534
-
19,480
3,902
8,260
8,181
8,033
15,998
13,125
Mills/kWh
4.29
5.90
5.08
3.62
3.65
3.67
4.45
2.93
-
5.57
2.79
2.36
2.34
2.30
4.57
1.87
a Power unit on-stream time, 7,000 hr/yr. Midwest plant location, 1978 revenue requirements. Investment and revenue requirements for disposal
, of flyash excluded.
These revenue requirements reflect capital investments shown in Table 1 (updated); byproduct credit and sludge fixation costs excluded.
-------
Table 3. DATA USED IN REVISION OF TABLES 11 AND 22 OF THE TVA PAPER
"COST COMPARISONS OF FLUE GAS DESULFURIZATION SYSTEMS"
Data for Revision of Capital Investment
Investment basis: 500-MW, new coal-fired unit burning coal with 12% ash,
30-*yr life, 127,500 hr of operation, only pumps spared, no bypass ducts,
no overtime, experienced design and construction team, 3-yr project, 1975-78;
1977 costs for scale, includes both flyash and S02 removal, excludes flyash
disposal, reheat to 175°F.
Material index for direct costs
Labor index for direct costs
Mid-1974
153.8
177.9
Mid-1977
220.4
197.0
$/ton
$/ton
$/ton
$/ton
$/ton
$/liter
$/yr
$/lb
4.00
18.50-24.00
155.00
15.00
52.00
1.65
-
2.00
Data for Revision of Annual Revenue Requirements
Raw materials Cost basis 1975
Limestone
Lime (varying with quantity)
Magnesium oxide
Coke
Soda ash
Catalyst (MgO, Cat-Ox)
Catalyst (sodium)
Antioxidant
Conversion costs
Labor: Operating, supervision
Analyses
Utilities
Process water (vs. quantity)
Fuel oil, No. 6
Fuel oil, No. 2
Natural gas
Heat credit, coal
Heat credit, oil
$/man-hr
$/man-hr
$/M gal
$/gal
$/Mcf
$/MM Btu
$/MM Btu
8.00
12.00
0.02-0.08
0.23
0.30
1.00
0.60
1.60
Steam
Coal
Coal
Oil
Oil
Electricity
Coal
Coal
Oil
Oil
1975
1978
1975
1978
1975
1978
1975
1978
$/M Ib
$/M Ib
$/M Ib
$/M Ib
$/kWh
$/kWh
$/kWh
$/kWh
0.80
1.50
1.50
2.75
0.011
0.028
0.019
0.041
0.70
1.40
1.40
2.65
0.010
0.027
0.018
0.040
1978
6.00
33.00-42.00
215.00
28.00
78.00
2.20
1.5 x 1975 cost
2.75
10.00
15.00
0.03-0.11
0.35
0.42
2.00
1.15 ($22.80/ton coal)
2.40
Year Cost basis 200 MW 500 MW 1,000 MW
0.60
1.30
1.30
2.55
0.009
0.026
0.017
0.039
86
-------
Table 4. REVENUE REQUIREMENT COMPARISONS (1978) OF FGD SYSTEMS
INCLUDING BYPRODUCT CREDITS AND SLUDGE FIXATION COSTS
Annual revenue requirements,
Process
Limestone
Lime
Magnesia
Cat-Ox
Sodium
Tons
byproduct
206,000d
175,000d
110,400e
137,400f
32,700^
As shown
in Table 2a
3.41
3.65
4.02
3.65
5.37
Includes
byproduct credit
-
-
3.23
3.04
4.99
mills /kWh
With sludge
fixation0
4.01
4.15
-
-
-
a New 500-MW, coal-fired unit, 3.5% S in coal, 7,000 hr/yr.
b 100% sulfuric acid at $25/ton, 80% acid at $16/ton, sulfur at
$45/long ton ($40.18/ton).
c Assuming sludge fixation service fee by contract treating sludge
in the utility's pond at $10/ton of 100% solids.
d 100% solids.
e 100% H2S04.
f 80% H2S04.
8 Elemental sulfur.
87
-------
These updated economics unfortunately do not reflect any recent
changes in process technology which, of course, should be taken into account.
In addition to sludge fixation, numerous other developments in lime and lime-
stone technology have taken place recently including (1) spray towers in
place of mobile-bed scrubbing devices, (2) series hold tanks, magnesium
addition and chloride ion control, pH and stoichiometry adjustments for in-
creasing limestone utilization, (3) mist eliminator wash cycles and hot air
injection or recycle reheat schemes to increase operating reliability, and
(4) sludge oxidation for more effective disposal. Furthermore, in spite of
the increased costs, some utilities have been leaning toward scrubber re-
dundancy and equipment sparing as means for improving system reliability.
Although the cost effects of these changes have not been explored
in updating the 1974 EPA-TVA cost study results, some of them have been
examined as part of a separate study which will be presented as part of this
paper.
For the three regenerable processes, there have not been as many
noteworthy developments as with lime-limestone; however, some new information
is available. For instance, data from the Boston Edison Company and Potomac
Electric Power Company (PEPCo) magnesia scrubbing demonstrations indicate the
process chemistry is workable as predicted, but some improvements are needed
in liquid-solids separation, drying, calcining and materials handling equip-
ment, and rubber-lined piping is definitely required. Since the piping in
the EPA-TVA estimates was rubber lined, the expected cost effects of the above
improvements would probably be limited to less than 10% over the 1975-78 pro-
jected costs.
Changes in the Wellman-Lord - Allied process primarily center- around
S02 regeneration where multiple-effect evaporation can be used to reduce
energy requirements at the expense of additional capital. Probably, a savings
of 5-6% in operating cost could be obtained with 2-4% increase in investment.
It is also anticipated that some measures may be taken to reduce or eliminate
the byproduct Na2S(>4 generated by the process; however, such process alter-
ations are yet under study and remain to be firmly applied.
Although the Wood River Cat-Ox demonstration has been plagued with
a variety of problems, many unrelated to the process itself, we are not aware
of any improvements or changes which would significantly alter the economics
as presented.
Once these and other regeneration process demonstrations are
carried to successful completion, there no doubt will be greater acceptance
by the utilities. The regeneration processes badly need a success story,
without any qualifications, to improve their image.
88
-------
EXPANDED EPA-TVA PROCESS ECONOMICS PROGRAM
In the months since the report "Detailed Cost Estimates for
Advanced Effluent Desulfurization Processes" was completed, TVA, under the
sponsorship of EPA, has begun a larger, expanded program in evaluating FGD
process economics which we hope will produce additional valuable results.
The phases of study are as follows:
Economics of lime-limestone scrubbing based on results of the
Shawnee test program
Byproduct marketing
Advanced process evaluations
Sludge disposal design and cost study
Energy optimization studies
Economics of Lime-Limestone Scrubbing Based on Results of the
Shawnee Test Program
Joint Bechtel/TVA Computer Study. Within the scope of the Shawnee
lime-limestone wet scrubbing test program, EPA is funding a joint study by
Bechtel Corporation and TVA to develop a computer program which will predict
investment and revenue requirements for various designs and/or operating con-
ditions. This project has been under way since early 1975. The program can
be used as a preliminary design and cost guide in choosing a lime-limestone
scrubbing system for a particular application (e.g. limestone vs. lime scrub-
bing utilizing TCA scrubbers). It can also be used to determine the effect
of a change of independent variables on the cost of a particular system
(e.g. 8 ft/sec vs. 12.5 ft/sec scrubber gas velocity). The model should not
be counted on, however, to compute the economics of a given system to a high
degree of accuracy.
Bechtel's major responsibility in the study is to provide design
equations for relating equipment size or capacity to the input process de-
sign variables. TVA is responsible for (1) the writing of the overall com-
puter program, (2) the generation of tables and/or equations relating
equipment cost to size or capacity, and (3) the estimation of total invest-
ment and revenue requirements, based upon the equipment costs and other
economic inputs.
The overall program is not yet complete; however, Bechtel has
completed the initial effort for specifying material flow rates and equipment
sizes, and TVA has developed the methods for costing the equipment. The
initial version of the overall program is expected to be complete by mid-1976.
Upon completion of the computer program, a user will be capable of comparing
the effect of various design conditions on relative economics of lime-limestone
systems for coal-fired power units ranging from 100 to 1300 MW. As a further
introduction to the program, the major input requirements for the joint
Bechtel/TVA computer program are listed below.
89
-------
1. Power unit size, MW
2. Boiler heat rate, Btu/kWh
3a. Coal analysis, wt % as fired
Carbon
Hydrogen
Oxygen
Nitrogen
Sulfur
Chloride
Ash
Moisture
3b. Heating value of coal, Btu/lb
4. Percent of sulfur emitted as S02 or 803 overhead
5. Percent of ash emitted overhead
6. Excess air to boiler, %
7. New versus retrofit installation
8. Limestone or lime chemical and physical properties
9. TCA scrubber operating parameters
Number of scrubbers
Number of beds
Scrubber gas velocity
Liquid to gas ratio
S02 removal, %
Stoichiometry
10. Redundancy desired
Number of spare scrubbing trains
11. Mist eliminator system options
With or without wash tray
12. Effluent hold tank definition
One, two, or three stirred tanks (series)
Slurry residence time, min (total)
13. Reheat (in-line steam reheater)
Outlet flue gas temperature, °F
14. Electrostatic precipitator or mechanical collector
Ash removal efficiency upstream of S02 scrubbers
15. Sludge disposal method
Onsite ponding: Distance between disposal site and scrubber systa
Weight percent solids in settled sludge
Fixation: Method or costs
90
-------
16. Power unit load factors
17. Remaining life of plant
18. Economic factors
Construction cost indices
Raw material and utility unit costs
The major outputs of the joint Bechtel/TVA computer program are
listed below.
1. Conceptual design information
Material balance
Number of trains
Equipment size, general specifications, and operating conditions
2. Capital investment
Equipment costs
Investment breakdown by area (feed preparation, scrubbing, disposal)
Investment breakdown by item (equipment, piping, ducts, structure,
electrical, instrumentation, etc.)
3. Annual revenue requirement
Direct costs: Raw materials
Labor
Utilities
Maintenance
Indirect costs: Capital charges
Overheads
Total annual revenue requirement
4. Lifetime revenue requirement
Total
Average unit
Levelized unit
EPA-TVA Limestone Process Economic Sensitivity Study. In April
1975 TVA completed a sensitivity study for EPA showing the effect of several
selected operating variables on the costs of limestone scrubbing. As an
example of the potential use of the computer program upon completion, the
results of selected design options considered in the April 1975 sensitivity
study have been revised manually to reflect more accurately the equipment
definition for the computer study. The updated results are shown in Table 5
as ratios to total investment and total revenue requirements of the base case.
In the computer study, results will be calculated to give a total project
cost rather than as a ratio to a base case.
91
-------
Table 5. LIMESTONE SLURRY PROCESS
PROJECTED INVESTMENT AND REVENUE REQUIREMENT RATIOS3
Total Total annual
investment revenue requirement
Case
1
2
3
4
5a
5b
6a
6b
7
8
9
10
11
12
13a
13b
14a
Ub
15
16
Identification
Base case
12.5 ft/sec gas velocity
12.5 ft/sec gas velocity with wash trays
MgO addition
Hypalon-lined pond
Pond without clay lining
Sludge fixation service fee - $5/wet ton (50% solids)
Sludge fixation service fee - $10/wet ton (50% solids)
Slurry oxidation - low pH operation with series
hold tanks
Slurry oxidation - MgO addition
Benzole acid addition
Series hold tanks
Low pH operation
Low pH operation with series hold tanks
Hot air injection reheat - 100 F
Hot air injection reheat - 50 F
Flue gas recirculation reheat - 100 F
Flue gas recirculation reheat - 50 F
In-line reheat - 50 F
In-line reheat - 100°F - entraimnent level - 0.67%
ratio
1.000
0.915
0.931
0.911
1.125
0.981
0.839b
0.839b
1.033
1.030
0.990
0.997
0.990
1.006
1.080
1.022
1.145
1.037
0.984
1.001
ratio
1.000
0.935
0.975
0.974
1.070
0.989
1.099
1.287
1.140
1.177
1.008
0.990
0.994
1.001
1.209
0.989
1.100
0.958
0.920
1.032
3 500-MW new coal-fired power unit, 3.5% S in fuel, 7,000 hr/yr.
85% SO removal, mid-1977 investment, mid-1978 revenue requirements.
Assumes service contractor provides investment for sludge transportation, fixation, and
disposal. Includes only cake filtration and loading facilities - pond excluded.
-------
Base Scrubbing System Incorporated in the Joint Bechtel/TVA
Computer Program and the Economic Sensitivity Study. The base system re-
flected in the computer program and sensitivity study utilizes a forced-draft,
dry fan system upstream of the S(>2 scrubbers. Costs include a low-efficiency
mechanical collector upstream of the fans to insure removal of the large ash
particles in the gas thereby preventing erosion of the fan blades. A plenum
is used for separation of the power plant ducts from the scrubber ducts,
allowing the number of scrubber trains to be a variable, independent of the
number of power plant ducts. Although separate fans are provided on each
side of the plenum to insure satisfactory gas distribution, only the scrubber
fan costs are included within the estimates. The SC>2 scrubber is designed
with a presaturator compartment for partial humidification and cooling of the
gas upstream of the first bed. The number of scrubber beds and depth of
spheres per bed are inputs.
An in-line steam reheater is located downstream of the scrubber in
the vertical section of duct. The reheater is designed with Inconel 625
tubes for reheating the gas to 150°F. All reheater tubes in contact with
the gas above 150°F are Corten. The gas is reheated to a final temperature
of 225°F.
With the exception of the "fixed sludge" variations, each case is
designed and sized for onsite pond disposal of wet collected flyash in
addition to the SC>2 byproducts. However, costs for transporting and dis-
posing the ash removed by the mechanical collectors are excluded. The
distance between the scrubbers and the disposal pond for each case is 1 mi.
The concentration of solids in the settled sludge is assumed to be 40% by
weight for all cases except those designed for high oxidation of sulfites to
sulfates. The concentration of settled solids in the pond for the high oxida-
tion cases is assumed to be 60% by weight. Disposal lines (spares included)
to the pond are rubber-lined carbon steel. Supernate return lines are carbon
steel. Ponds are sized for a total of 127,500 hr of operation over the life
of the plant, corresponding to an average capacity factor of 48.5% over a
30-yr plant life.
93
-------
A detailed listing of the cases and premises for these updated
costs and the initial computer program is shown below.
Definition of Cases for Updated TVA Economic Sensitivity Analyses.
Base Case
Case 1: Limestone additive
Megawatts = 500 (4 modules)
Sulfur in coal = 3.3 wt %; ash in coal = 12 wt %
TCA with 3 stages (4 grids), 5 in. spheres/stage
Scrubber pressure drop =4.2 in. H20 (excluding mist
eliminator)
Chevron mist eliminator pressure drop =0.2 in. H20
L/G - 75 gal/Mcf
Liquor rate = 36 gpm/ft^
Gas velocity = 8 ft/sec
Limestone utilization = 65%
S02 removal = 85%
Particulate removal = 99.5%
Solids in recirculated slurry = 15 wt %
Scrubber inlet liquor pH = 5.8
Effluent residence time (single tank) = 12 min
Pond (clay-lined) for sludge disposal
In-line steam (500 psig pressure) reheater, 100°F reheat
FD fan, with partial coarse ash cleanup, 33% efficient
mechanical collector
Scrubber inlet particulate loading = 2-3 gr/scf
Velocity Effect
Case 2: Same as Case 1, except:
Gas velocity = 12.5 ft/sec
Liquor rate = 33 gpm/ft^
L/G = 44 gal/Mcf
Scrubber pressure drop =7.2 in. 1^0 (excluding mist
eliminator)
Chevron mist eliminator pressure drop = 0.3 in. H20
Case 3: Same as Case 1, except:
Gas velocity =12.5 ft/sec
Liquor rate - 33 gpm/ft2
L/G = 44 gal/Mcf
Wash tray plus chevron mist eliminator
Scrubber pressure drop = 7.2 in. t^O (excluding mist
eliminator and wash tray)
Chevron mist eliminator pressure drop =0.3 in.
Wash tray pressure drop = 2.7 in.
94
-------
MgO Addition Effect
Case 4: Same as Case 1, except:
Liquor magnesium ion concentration = 10,000 ppm
Effluent residence time = 5 min
L/G = 40 gal/Mcf
Liquor rate = 19.2 gpm/ft2
Scrubber pressure drop =3.1 in. 1^0 (excluding mist
eliminator)
Alkali Utilization/Sludge Disposal Effects
Case 5a: Same as Case 1, except Hypalon-lined pond
Case 5b: Same as Case 1, except pond without clay lining
Case 6a: Same as Case 1, except no pond; sludge fixed and disposed
under service contract of $5/wet ton (50% solids)
Case 6b: Same as Case 6a, except no pond; sludge fixed and disposed
under service contract of $10/wet ton (50% solids)
Case 7: Same as Case 1, except:
12 in. spheres/stage
Scrubber pressure drop = 7.2 in. H20 (excluding mist
eliminator)
Limestone utilization = 90%
Scrubber inlet liquor pH = 5.2
Three series hold tanks (4 min each)
Complete oxidation of slurry bleed stream (pH -4.5) from the
first hold tank, using air at 5 atm and stoichiometry of
5 moles 02/mole S02 absorbed (similar to JECCO mode).
Case 8: Same as Case 4, except complete oxidation of slurry bleed
stream (from the single hold tank) using air at 3 atm and
stoichiometry of 5 moles 02/mole S02 absorbed (similar to
Kellogg mode). Use 1 hr residence time in the oxidizer.
Case 9: Same as Case 1, except:
Organic additive
Limestone utilization = 80%
Case 10: Same as Case 1, except:
Three series hold tanks (4 min each)
Limestone utilization = 75%
Case 11: Same as Case 1, except:
Scrubber inlet liquor pH = 5.2
Limestone utilization = 80%
12 in. spheres/stage
Scrubber pressure drop = 7.2 in. H20 (excluding mist
eliminator)
95
-------
Case 12: Same as Case 7, except no oxidation of slurry bleed stream.
Reheater Effect
Case 13a: Same as Case 1, except:
External steam reheater with air - 100°F
Case 13b: Same as Case 13a, except - 50°F
Case 14a: Same as Case 1, except:
External steam reheater with flue gas recycle - 100°F
Case 14b: Same as Case 14a, except - 50°F
Case 15: Same as Case 1, except - 50°F
Case 16: Same as Case 1, except entrainment level - 0.67% of
scrubber outlet flue gas rate
Byproduct Marketing
The second most advanced project in the overall EPA-TVA FGD economics
program is an intensive study of marketing byproducts from FGD systems. By-
products from both nonregenerable and regenerable processes are being evaluated
for potential market volume, location, and price.
The initial efforts in this area were undertaken in early 1973 when
a hypothetical study was made of marketing sulfuric acid which TVA could con-
ceivably produce from its coal-fired power plants. Using a computeriEed trans-
portation model a report entitled "Marketing I^SO,/, from S(>2 Abatement Sources—
The TVA Hypothesis" was published in December 1973 indicating that approximately
2 million tons per year of abatement acid could be produced. Under the market
conditions prevailing at that time, a net sales revenue of $5-9/ton was indi-
cated. With recent price increases, no doubt higher revenues could now be
expected.
In June 1974 TVA prepared a second byproduct marketing study for EPA
entitled "Preliminary Feasibility Study of Calcium-Sulfur Sludge Utilization
in the Wallboard Industry." Although limited in scope and very brief, this
study indicated the possible savings to be derived from marketing sludge as
gypsum.
With these two projects as pilots, agreement was reached in 1974 to
expand the sulfuric acid analyses to cover both sulfuric acid and elemental
sulfur for all the utilities in the United States and to carry out a more com-
prehensive study of gypsum markets. The computer program used for the TVA
hypothetical assessment was to be modified and expanded considerably for this
effort.
96
-------
During the early phases of the expanded acid study, it became
apparent that a sophisticated, useful computer program applicable to any by-
product was being developed an2 control.
Both double-alkali processes which produce sludge and regenerable processes
producing sulfur are already under study.
In support of the EPA demonstration programs, preliminary, rough
assessments (new 500-MW, coal-fired units, 3.5% S in coal) were made of the
Atomics International's aqueous carbonate, UOP - Shell's copper oxide, Air
Products - Catalytic's ammonia/IFP, and Consol's potassium scrubbing pro-
cesses. These processes all produce elemental sulfur. Preliminary results
(yet to be confirmed by detailed study) indicate a wide range of investment
($87 to $160/kW for 1975-78 project) and total annual revenue requirements
(3.7 to 7.5 mills/kWh for 1978 startup). If $45/long ton of sulfur is taken as
•byproduct credit, the net revenue requirements range from 3.3 to 7.1 mills/
kWh.
In addition to the regenerable processes, three double-alkali sys-
tems— Envirotech, CEA - ADL, and FMC—have been studied. The total investment
(including sludge pond) and revenue requirements of the three systems are
relatively close, averaging about $71/kW investment and 3.6 mills/kWh annual
revenue requirement. Additional refinement of vendor data will be necessary
before true comparative costs can be derived.
97
-------
Thus far, in all processes evaluated, capital charges (excluding
maintenance) constitutes the largest fraction of annual revenue requirements
ranging from about 40-43% for the double-alkali processes to about 45-60% for
the regenerable processes. In the double-alkali processes, raw materials
(primarily lime) require the second greatest expenditure with energy third.
This is reversed with regenerable processes in that they require more energy
(fuels, electricity, and steam) than raw materials. (This can be misleading
since components such as naphtha, coke, coal, and fuel oil could easily be
classified as raw materials or energy.) Operating labor cost for all the
processes is less than 3% of total revenue requirements.
Plans now call for more intensive evaluation of the most promising
systems. Detailed work is already under way on the Bureau of Mines' citrate
process for which design, investment, and revenue requirement data will be
developed along the lines of the earlier EPA-TVA studies. After review by
the process developers for accuracy, reports will be published as with previous
evaluations.
Sludge Disposal Design and Cost Study
Another area of work to be covered in the process economics program
is a detailed conceptual design and cost study of sludge disposal processes.
In the past the lack of detailed design and cost data for this area has re-
duced the value of our process economic assessment effort. Included will be
cases covering several pond design and liner options, fixation processes such
as IUCS, Dravo, and Chemfix, and various sludge dewatering techniques. In
addition, accurate design and cost data will be developed for combined and
separated flyash - scrubber sludge disposal systems. This work is just getting
under way and the results will also be published.
Energy Optimization Studies
The final phase of economic activity covers a careful investigation
of SC>2 removal process energy requirements and is directed toward reducing or
minimizing process energy needs. An effort will be made to obtain data from
actual scrubber installations to compare with theoretical needs. In addition,
key unit operations will be analyzed toward finding alternate routes to lower
energy consumption. Work on this project is to be started soon.
ECONOMICS FOR THE TVA OFFICE OF POWER
In addition to the work being performed for EPA, the OACD studies
staff has continued to support the TVA Office of Power with specific energy-
environmental-related process evaluations. In the area of FGD, evaluations
are now under way on two promising Japanese double-alkali processes—
Kureha and Dowa. The Kureha system uses a sodium acetate scrubbing solution
to react with limestone whereas the Dowa process uses a soluble aluminum
sulfate - oxide complex in water to absorb S02 for further reaction with lime-
stone. Recent visitors to Japan have been impressed with these two processes
which produce gypsum for sale or disposal by piling.
98
-------
Results thus far indicate these two processes to be competitive
with double-alkali systems under development in the United States. If
sludge fixation or other extra processing measures (reduction of soluble
salt losses, oxidation of sludge) are required, these processes will have
strong advantages.
Also, during the past year, OACD has taken part in other Office
of Power energy process evaluations such as solvent refined coal, fixed-bed
coal gasification, NOX removal studies, and fluidized-bed combustion. Some
of these results have been reported previously. These activities have
enabled us to make comparisons of the various power plant environmental con-
trol options with perspective and consistency.
As can be seen, TVA has been and continues to be involved in a
number of studies on FGD system economics. In meeting these and future
responsibilities, we are striving to keep our quality high so that repre-
sentative process comparisons can continue to be made to guide the industry
in making key decisions.
99
-------
STATUS OF THE EPRI FLUE GAS DESULFURIZATION
DEVELOPMENT PROGRAM
Lawrence W. Nannen and Kurt E. Yeager
Electric Power Research Institute
Palo Alto, California
ABSTRACT
This paper summarizes the present status of EPRI's program in flue
gas desulfurization for both lime/limestone and advanced regenerable
processes. Emphasis is given to problem areas in lime-limestone scrubbing
since it offers the most widely accepted near-term solution to positive
S02 emission control. Specific EPRI projects are discussed and general
guidelines for reliable scrubber implementation are offered.
101
-------
STATUS OF THE EPRI FLUE GAS DESULFURIZATION
DEVELOPMENT PROGRAM
by
Lawrence W. Nannen
and
Kurt E. Yeager
Electric Power Research Institute
Palo Alto, California
The Electric Power Research Institute began operations in 1973 for the
purpose of expanding electric energy research and development under the
sponsorship of the nation's utility industry. EPRI is presently funded
by over 500 member organizations, public, private, and cooperative.
EPRI's overall goal is to develop a broad, coordinated, advanced technology
program for improving electric power production, transmission, distribution,
and utilization in an environmentally acceptable manner. The primary
areas of EPRI's research are fossil fuel and advanced systems; nuclear
power; power transmission and distribution; and energy systems, environment,
and conservation.
Projects of high priority at EPRI include those which develop technological
options to permit continued utilization of coal for power production in
both new and retrofit applications within environmental constraints. A
major current issue associated with all fossil fuel combustion, particularly
coal, is the control of sulfur oxide emissions.
Current EPRI Objectives in Flue Gas Desulfurization
The present research and development projects dealing with SO2 control
within EPRI's Environmental Control and Combustion Program are oriented
to providing electric utilities with reliable, cost-effective technology
for removing sulfur oxides from coal-fired steam boilers. The goals of
these projects are to ultimately develop a reliable commercial design
basis for both throwaway scrubbing systems and regenerable FGD technology
so that utilities can apply such systems if they are required.
102
-------
LIME/LIMESTONE SCRUBBING
Since the technology for lime/limestone scrubbing is generally considered
to be the most advanced for present commercial applications to utility
boilers, and since the bulk of utility commitment is towards this tech-
nology (40,000 MW and $3 billion by 1980), the obvious near-term objective
of EPRI is to assure that the technical issues are resolved to allow
reliable, efficient operation of these systems. A large amount of
operating experience now exists for lime/limestone scrubbing — some
good, some bad, and most poorly understood. Therefore, a major task is
to collect the available data, sort out areas in which the technology
needs improvement, and develop a methodology for designing these processes
on a site-specific basis.
Table 1 lists the major problem areas continuing to face the utility
industry in integrating lime/limestone scrubbing with power production.
Table 1
Major Problem Areas in Lime/Limestone Scrubbing
* Mist Elimination
* Reheat
* Materials of Construction
* Process Chemistry and System Design for Site Specific
Applications
* Waste Handling and Disposal
As is the case with any complex process, one must maintain the awareness
that each problem area is interrelated with another and must be addressed
with the total system in mind.
Mist Eliminator
The function of a mist eliminator in any wet scrubbing system is to
prevent the carryover of aqueous aerosol droplets downstream from the
absorber to prevent plugging with slurry solids and corrosion of reheaters,
scrubber I.D. fans, and flue gas duct work. The design of the mist
eliminator system must consider the following variables:
Orientation. The demister may be placed horizontally in a vertical gas
duct or vertically in a horizontal gas duct. It may also be inclined to
assist drainage during washing. It appears that a vertical demister
configuration (horizontal gas flow) offers advantages of superior drainage,
greater washing flexibility, and better adaptability to closed-loop
demister wash water recycle.
103
-------
Flow Distribution and Velocity. Most mist eliminators used in commercial
FGD scrubbing systems are of the zig-zag "Chevron" baffle design.
Proper flow distribution and velocity of the wet gas through the demister
is essential for proper operation and efficiency. The radial vane
collector shows promise in improving gas distribution.
Washing. Reliable operation of mist eliminators in lime/limestone
scrubbing systems is complicated by the normally high solids content of
the scrubbing liquor. These liquids can adhere to the demister surface,
forming a soft mud deposit or a hard chemical scale. A large amount of
development work is presently underway to design a washing system to
minimize solids buildup on the demister.
Tie-in with the Overall System. If the scrubber system is designed for
closed-loop operation (no water discharge), the quantity of fresh water
available to the system is limited to the water lost by saturation of
the flue gas and water remaining with the waste solids after final
disposal. Most experience to date clearly indicates that the demister
wash water should be of the best quality, with the minimum amount of
dissolved calcium, to avoid scaling. There is also a greater tendency
for the demister to plug when limestone is used rather than lime as the
absorbent. This is caused by the normally poorer utilization of the
limestone which allows excess unreacted calcium to be entrained into the
demister. The use of series hold tanks and low inlet slurry pH (between
5.0 and 5.5} can improve limestone utilization; however, some reduction
in SO2 removal efficiency may occur at low scrubbing pH.
Reheater
The function of a reheater is to warm the flue gas above its wet bulb
temperature of 125°F after leaving the scrubber to a temperature normally
between 175°F and 200°F. The warmer gas is then less likely to condense
and corrode fans, duct work, and gas dampers. Reheat also increases
plume rise and dispersion of the effluent to minimize ground-level
concentrations of not only SO2 but also particulates and nitrogen oxides.
Degree of Reheat Needed. There is some question as to the degree of
reheat needed for site-specific meteorological conditions. There is no
legal requirement for reheat in present standards.
In-Line Steam Tube Reheat. The most common design of reheat is a
series of steam tubes inserted in the flue gas stream. Good reheater
operation then depends largely on proper mist eliminator performance.
However, corrosion is a major problem.
Indirect Reheat. A more costly, less efficient but very reliable design
is to blend heated air with the flue gas to attain the desired temperature.
This approach results in reduced maintenance costs.
104
-------
Hot Gas Bypass. Use of hot gas bypass is dependent on the efficiency of
the scrubber to remove a very high percentage of the SO2 in the processed
flue gas.
Fuel Combustion. The use of natural gas or fuel oil for direct combustion
irt the exit gas stream may be attractive in a few isolated cases, but is
a poor option for the future due to the risk of fuel shortages.
Materials of Construction
Lime/limestone process slurries contain abrasive solids (limestone, fly
ash, and gypsum) and corrosive dissolved species (chlorine, hydrogen,
and concentrated salts). A major design problem is selection of reasonable
cost materials of construction to minimize corrosion and erosion of
scrubber components.
Liquor Loop. Pumps, valves, piping, and spray nozzles are all affected
by process slurry conditions.
Flue Gas Handling. Fans, duct work, and the reheater must withstand the
corrosive flue gas environment.
Absorber Internals. Scrubber internals such as the mist eliminator,
stage partitions, trays, and packing must withstand both gas and liquor
contact.
Process Chemistry and System Design
The previously discussed problem areas were primarily hardware issues.
It is important to recognize that the poorest understanding of lime/limestone
scrubbing is in the process chemistry and its effect on overall system
design. We are still unable to confidently predict the chemical behavior
of a scrubber for each potential utility site without a time-consuming,
expensive pilot testing program at each site. Table 2 lists the important
site-specific variables which affect scrubber system design.
The major scrubber chemistry issues requiring further study are:
Degree of Sulfite to Sulfate Oxidation. How is oxidation promoted or
retarded in the scrubber? How is oxidation of the waste product achieved
to allow better settling and dewatering?
Effect of Fly Ash Chemistry. Fly ash provides surface area for possible
Chemical reactions. Also, alkaline fly ashes from Western coals can be
used beneficially in the scrubbing process.
Supersaturated vs. unsaturated Gypsum Chemistry. We do not fully understand
what conditions cause operation unsaturated with respect to calcium
sulfate.
105
-------
Table 2
Process Chemistry and System Design
Site-Specific Variables
Degree of SO2 removal required
Inlet SO_ concentration
Excess air
Degree of oxidation
Fly ash loading
Fly ash chemistry and solubility
Degree of closed-loop operation required
Dissolved solids in system
Tie-in with other plant waste water streams
Constraints on sludge disposal
Chemistry Affects System Design
Absorber type
Reaction tank
Liquid-to-gas ratio
Reactant type
Process control
Materials selection
Sludge composition and handling
106
-------
Effect, of Magnesium and Chlorine on scrubber chemistry.
Use of Lime vs. Limestone. Limestone is the most economical, energy
efficient reagent but in many cases complicates the scrubber chemistry.
More work is needed to understand how to increase limestone utilization
while maintaining high SC>2 removals.
Use Cooling Tower Slowdown for Closed Loop Operation. The scrubber has
the potential to act as a "garbage disposal" for the power plant. Water
supplies can be conserved if scrubbers can feed off other plant waste
streams.
Waste Handling and Disposal
Handling and disposal of lime/limestone scrubber sludge is a complex
site-specific problem for which usable guidelines are not available.
Scrubber wastes consist of three general types of material: fly ash,
calcium sulfate/sulfite salts, and scrubbing liquor associated with the
partially dewatered sludge. Each type of material has different physical
and chemical characteristics. For example, trace elements are found
almost exclusively in the fly ash, the calcium sulfite has very poor
physical properties resulting in inadequate dewatering and structural
stability features, and the liquor contains concentrated dissolved salts
produced from the scrubbing process. Disposal and treatment methods
have started to evolve without a clear understanding of each sludge
component; the problems unique to one general component have been confused
to be characteristic of the total sludge. Problems such as trace element
leachability, sludge fixation, sludge dewatering, and beneficial use of
scrubber wastes cannot properly be addressed until the characteristics
of each sludge component are understood.
EPRI Task Areas for Throwaway SO Scrubbing
The following discussion presents the methodology and organization of
the R&D tasks in EPRI's throwaway S02 program.
1. Evaluate Utility Operating Experience (1974-75)
EPRI is funding the Battelle Stack Gas Coordination Center. Battelle
has completed a state-of-the'-art study for EPRI summarizing the commercial
SO2 experience in the U.S. and recommending further areas of research.
EPRI has also formed a utility working group to advise the EPA and TVA
on the conduct of the Shawnee test program.
2. Unit Analyses of Scrubber and Disposal Hardware (1975-76)
Battelle is continuing their effort to evaluate scrubber problem areas
with a series of "unit analyses" on mist elimination, reheat, and cost
variation in scrubber bids.
107
-------
A new contract is under negotiation with Southern California Edison and
Steams-Roger to study improved process control capabilities for scrubbers
and the instrumentation available to measure control variables.
Proposed new projects to evaluate sludge handling techniques and to
recommend pump, fan, and valve materials and design are planned.
EPRI«is also supporting the Southern Services testing at Plant Sholz in
Florida which is operating three 20 MW pilot plants: double alkali,
dilute acid Chiyoda, and Foster Wheeler carbon adsorption.
3. Develop and Test Hardware and Chemistry (1976-77)
The Tennessee Valley Authority has been contracted by EPRI to study four
tasks: a) horizontal 1 MW pilot scrubber for high sulfur coal, b) improved
methods of reheat, c) corrosion/erosion materials testing, and d) characteri-
zation of scrubber sludge as a function of scrubber operating conditions.
A contract is under negotiation'with Radian corporation to demonstrate a
novel concept in scrubber process control.
New projects include a study to characterize low sulfur/alkaline ash
scrubbing chemistry for Western coals and an investigation into the
oxidation chemistry of sulfite to sulfate.
4. Diagnostic Capability (1977-80)
It is planned that a large portion of EPRI's efforts in the future will
be centered around diagnostic testing of specific fuels in several
prototype scrubbers situated around the country. This service will
benefit the power industry by providing a means to demonstrate overall
scrubber operation and behavior before committing to a specific design.
Diagnostic sites will be developed for both Eastern high sulfur coal and
Western low sulfur coal.
Figure 1 describes the current EPRI SOx Control Program and illustrates
the interrelation and timing of specific projects for both lime/limestone
scrubbing and regenerable FGD technology.
108
-------
OBJECTIVE
FIGURE 1
SOX CONTROL PROGRAM
PROJECT DESCRIPTION
Develop Reliable
Design Basis for
Lime/Limestone
Scrubbing
1. Evaluate Utility Operating Experience
2. Unit Studies-Demister, Reheat, Cost Variation
3. 1 MW Tests-Reheat, Corrosion, Horizontal Scrubber
4. Evaluate FGD Process Control Capability
5. Design Guidelines for Lime/Limestons Scrubbing
6. Low Sulfur/Alkaline Ash Scrubbing Characterization
7. Evaluation of Sludge Dewatering Processes
8. Sludge Composition & Leachability
9. Sludge Fixation Chemistry Guidelines
10. Chemistry Modifications To Improve Reliability & Cost
11. Hardware Modification To Improve Reliability & Cost
12. Diagnostic Support - Eastern Coal
13. Diagnostic Support - Wstern Coal
Develop Viable
Begenerable FGD
Technology
14. Comparative Evaluation of Regenerable FGD Processes
15. Comparative Design Of Advanced Processes
16. Co-Sponsored Pilot (10-40MW) Process Construction
& Operation
17. 20 MW Pilot Test Of Chiyoda, Double Alkali,
Bergbau Processes
18. Modification To Existing Processes To Improve
Reliability & Cost
19 o Test & Evaluation Of Commercial Scale (100MW+)
Processes
109
-------
FIGURE 1 (continued)
SOX CONTROL PROGRAM
75
i A
, r,.
76
J 2 L
L_
! 14
t
1 n
700
4
n
T
1 6
1 7
77
MIL
_ /
L
A
f.
L
f^ /
_
1 8 h
h
1 15
y
— , -U
r?
L-*|
4400
r
78
ESTONES
^ - Def in
^v - Compl
\ - Updat
^y - Decis
79
it ion of K
ete State-
e State-Of
ion on Pro
10 LV — '
. , ^ _ _ ^
I. 12
1 13
^. 16
18
4900
r^l
Jt ,.
II- . 1
($ x
3600
110
a_
1000)
2500
80
ey Problem
Of-The-Art
-The-Art G
cesses To
81 | 82
Areas
Guideline
uide lines
Be Constru
i
i
2000
(5000)
2000
(5000)
1
s Document
Document
=ted
^J
1
2000
(5000)
-------
ADVANCED REGENERABLE PROCESSES
Recognizing the need to develop desulfurization processes which 1) do
not create a solid waste disposal problem, 2) are not dependent on
limestone raw materials, 3) have the possibility of being less sensitive
to site-specific design variables, and 4) can produce a useable byproduct.
EPRI is active in the evaluation of advanced regenerable processes. The
objective of a current contract with Radian Corporation is to compare
and evaluate a large number of advanced processes on technical grounds
(material and energy balances) and to recommend further areas for process
development.
EPRI's evaluation program is considering three broad areas of regenerable
process development:
1) Modifications to commercially available regenerable FGD processes
such as Wellman-Lord, magnesium oxide scrubbing, and CatOx to improve
performance, reliability and cost.
2) Development of a regeneration capability for calcium sulfite/
sulfate throwaway FGD processes. In certain circumstances, particularly
as a retrofit to lime/limestone scrubbing processes in the future, it
may be possible to convert the sludge into sulfur or acid or to convert
calcium based systems to magnesium oxide regenerable processes.
3) Development of new regenerable processes. A new EPRI project is
planned for 1976 which will provide a competitive site-specific design
evaluation for processes in all three of the above categories, with the
possible selection of one or two new processes for pilot plant (10-
40 MW) construction and demonstration with a host utility.
Ill
-------
CONCLUSIONS
For the near term, the primary utility choice for SO2 control appears to
be throwaway scrubbing processes involving the use of alkaline reagents
and the production of a calcium sulfite/sulfate waste product. Many
utilities are reluctant to install these systems for three fundamental
reasons:
1) Scrubbers are entirely non-profit for utilities and are expensive
to install and operate. There is no guarantee that increased costs can
be passed on to the customer in a timely fashion.
2) There is a questionable cost/benefit relationship and many utilities
doubt the need to control SO2 emissions to the degree many regulations
require.
3) Early experiences with scrubbers were very poor and utilities have
been frightened by their foreign nature to conventional power plant
operation. Only recently have utilities recognized the fact that scrubbers
are complex developmental chemical processes and operated them as such.
Considering these difficulties and the lack of positive incentive for
the utility industry to purchase and operate chemical processes which
are barely developed, scrubber technology has done well to evolve as far
as it has.
The following "challenges to industry" are given for the purpose of
maximizing the reliability of any scrubber installation in the utility
industry:
1) Utilities must assume responsibility to make the scrubber system
work. If the decision is made to install scrubbers, the utility must
accept the system as a developmental prototype and provide for an engineering
and maintenance intensive installation. Provided with a proper design
basis, the utility must have its own qualified staff of chemists as well
as mechanical and chemical engineers to work with its architect/engineer
and selected vendor in building the best system to meet site-specific
conditions. Process guarantees and fixed-cost contracts are not sufficient
to assure reliable, cost-effective scrubber operation. Plant personnel
must give the scrubber operating and maintenance priority equal to all
other power station systems.
2) Utilities should encourage vendors to achieve a mechanical "hybrid"
system based on the extensive body of available operational experience,
which incorporates the best components (i.e., absorber, mist eliminator,
reheater, process equipment) of all presently available scrubber approaches.
The goal of this effort is to standardize the scrubber hardware for low
sulfur and high sulfur coals such that development priorities may be
placed on chemical design.
112
-------
Prom what we know today, some fundamental considerations for maximizing
scrubber reliability include the following:
a) The SC>2 absorber should contain a minimum of internals to
avoid plugging and scaling. Where particulates are to be removed
in the scrubbing system, a two stage venturi/spray tower combination
is recommended.
b) A vertical mist eliminator (horizontal gas flow) offers advantages
of superior drainage, greater washing flexibility, and better
adaptability to closed-loop demister wash water recycle. Horizontal
mist eliminators appear to work satisfactorily for low sulfur cqal
and for high sulfur scrubbing with lime.
c) The most reliable, low maintenance reheater design appears to
be blending heated ambient air with the flue gas. This approach is
more capital intensive and less energy efficient, but is very
reliable.
d) Scrubber booster fans appear to operate more reliably when
they are dry. Forced draft fans and induced draft fans installed
downstream from the reheater have encountered minimum corrosion
problems.
e) Spare modules and backup components improve the ability of
scrubbers to be cleaned and serviced while on-line.
3) Concentrate on scrubber process chemistry for site-specific design.
Capability to predict chemical behavior and performance reliability
before the scrubber is built is needed.
113
-------
NON-REGENERABLE PROCESSES SESSION
Chairman: Michael A. Maxwell
Chief, Emissions/Effluent Technology Branch
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
115
-------
IERL-RTP SCRUBBER STUDIES RELATED TO FORCED OXIDATION
Robert H. Borgwardt
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ABSTRACT
Small scale experiments indicate that forced oxidation of FGD
scrubber sludge to gypsum is a feasible option for U.S. systems operating
with high-sulfur Eastern coals containing chloride. Forced oxidation
was carried out as an integral part of the first stage of a two stage
system, by air sparging the effluent hold tank of the first stage at
pH 4.5. Under these conditions complete limestone utilization was
achieved, together with total oxidation of the calcium sulfite at
atmospheric pressure. The performance of the oxidizer was limited
by mass transfer of 0_ from the air bubbles to the liquor, consequently
a tower containing a 5.5-m depth of slurry was used to maximize oxygen
transfer efficiency. Complete oxidation of the sludge was thus obtained
at an air stoichiometry of 2.6, without catalysts and without mechanical
atomizers. The oxidized slurry settled at a rate ten times faster
than that of sulfite slurries at normal oxidation levels. A settled
sludge of increased density (60% solids) also resulted, a factor
important to the environmental and economic aspects of sludge disposal.
The combined effects of high utilization and oxidation can yield a 35%
reduction in the amount of total wastes produced when the scrubber
is used for fly ash collection. An important objective of these tests,
to produce a sludge that is filterable to 80% solids, was not attained.
This result, together with the observation that the settling properties
can in some cases be dominated by the fly ash, favor the dry collection
of fly ash ahead of the scrubber as an alternative method of reducing
total waste.
117
-------
IERL-RTP SCRUBBER STUDIES RELATED TO FORCED OXIDATION
INTRODUCTION
A typical 600-MW power plant, burning Eastern (3.5Z S) coal, will
generate 15 short tons of SO_ per hour. When this amount of SO- is
reacted with limestone to form an environmentally acceptable waste
product, it yields as much as 100 tons of wet sludge per hour, exclusive
of fly ash. The environmental and economic aspects of sludge disposal
in quantities of this magnitude were cited at the previous FGD symposium
as the major problem in application of throwaway processes. In view of
the current preference on the part of the electric utility industry for
non-regenerable FGD systems—in spite of the sludge problem—one may assume
that the overall economics justify the expense and effort required to deal
with it. EPA's assessment of the situation (Princiotta, 1975) pointed to
the improvement of limestone utilization as one potential method of
reducing the amount of sludge produced. It is also known that the single
factor most influencing the amount of waste produced, per unit of S00
scrubbed, is the dewatering characteristics of the solid product. This
property has been tentatively related to the degree of oxidation (Rossoff,
1974), by comparison of sludges produced in various full scale scrubbers.
Experiments at EPA's pilot plant at Research Triangle Park have been
focused on this problem during the past year, both from the utilization
and the oxidation standpoints. Increasing the utilization is, of course,
an important objective for its own merits due to the effect on operating
cost and the effect on mist eliminator performance (reduced fouling) . High
utilization is also considered, in our approach, to be a necessary condition
for efficiently carry ing out the oxidation step; i.e., the problems of
achieving maximum utilization and oxidation are related. The ultimate
objective was therefore to eliminate unreacted limestone as well as
unoxidized calcium sulfite from the product solids. The purpose of this
paper is to evaluate the prospects for achieving this objective with
scrubbers operating at conditions that correspond to the use of high sulfur
coal.
118
-------
BACKGROUND
The feasibility of converting calcium sulfite scrubber slurry to
gypsum was first indicated by I.C.I, in 1935; their original process
design included an oxidation tower. That design, in which the oxidation
step was carried out within the primary scrubber loop, was not efficient
with respect to the amount of air required. Mitsubishi Heavy Industries
reported (Uno, 1971) several improvements in the state of the art at the
Second EPA Limestone Scrubbing Symposium. The MHI Lime/Gypsum Process
employs two-stage scrubbing to adjust the pH to 4 - 4.5, followed by
pressurized (3.8 - 4,8 atm) aeration using mechanically driven atomizers
designed by the Japan Engineering & Consulting Co. (JECCO). The JECCO
oxidizer, operating with catalysts and low pH, has been successfully applied
in Japan to produce pure gypsum suitable for many industrial and construction
uses. The suitability of this process to the operating conditions of U.S.
scrubbers, which have both fly ash and chloride present in the slurry, has
not been demonstrated. The experimental study at RTF was initiated primarily
because of'questions regarding these factors.
The MHI report showed that the mass transfer efficiency of the oxidizer,
with respect to oxygen absorption, drops sharply at pH's greater than about
5.3 (with no Cl~ present). Fundamental kinetic studies conducted in the
Soviet Union (Gladkii, 1974) showed a maximum rate for the oxidation of
calcium sulfite slurries at pH 4.5. The HS03~ ion was identified as the
active species undergoing oxidation and the homogeneous reaction rate was
found to correlate directly with the concentration of that ion. Concurrent
vork carried out for EPA at Arthur D. Little, Inc. (Berkowitz, 1975) confirmed
the Soviet results: oxidation rate is proportional to the bisulfite concen-
tration in solution and, as a result, is an order of magnitude faster at PH 4
than at pH 7.
pH AND UTILIZATION
Achieving the low pH required for efficient oxidation of calcium sulfite
slurry necessarily means that the scrubber must operate at a high level of
limestone utilization. At equilibrium the slurry will have a PH of 6.4 as
119
-------
long as any unreacted limestone is present (Wen, 1975). The fact that pH's
*
this high are rarely observed in practice is a strong clue that the system
does not operate at equilibrium, even at normal limestone feed stroichiometry.
As stoichiometry is reduced the concentration of unreacted limestone in the
slurry is lowered and kinetic factors become totally dominant (recent tests
at Shawnee with the scrubber operating at low stoichiometry showed that
90 hours of stirring was required to raise the slurry pH from 5.7 to 6.1).
The loss of SC>2 removal efficiency under these conditions imposes a constraint
on both the minimum pH and the maximum utilization that can be obtained with
scrubbers of current design.
Techniques of optimizing the performance of limestone scrubbers at low
stoichiometry—a necessary condition for improving utilization—have been
examined at RTF and are reported elsewhere (Borgwardt, 1975) in detail. That
study showed the kinetics of limestone dissolution/SO? precipitation can be
accelerated by modifying the scrubber effluent hold tank to simulate tubular
flow. Three or four stirred tanks, of 2 - 4 min. residence time each, were
found suitable. The modified configuration (Figure 1) will achieve greater
completion of the hold tank reactions in any given total residence time than
a single stirred tank of the same residence time. It was shown that a single-
loop scrubber operating at 18 - 20 cm water pressure drop could thus achieve
about 90% utilization at 82% S02 removal efficiency. Ninety percent SO-
removal efficiency and 94% limestone utilization were obtained by increasing
the tower pressure drop (holdup) to 30 cm water. Under these conditions scrub-
ber effluent pH's of 4.5 and lower were achieved. The relationship between
pH, chloride, and the concentration of S02 in the scrubber liquor is shown in
Figure 2 as observed in the RTF pilot plant.
Attempts to oxidize the purge slurry of a system operating in the scrubber
configuration indicated by Figure 1 resulted in low conversion. Although the
initial pH was less than 4.5, it increased during aeration to 6.4. It was
evident that the unreacted limestone in the slurry was sufficient, even at
utilizations of 90 - 94%, to slowly increase the pH so that efficient oxidation
rates (and CaSO-'^H-O dissolution rate) could not be maintained.
120
-------
FLUE GAS
IN
OUT
A"
nrrmo
iiiiii
LIMESTONE FEED
SCRUBBER EFFLUENT HOLD TANK
SLUDGE
Figure 1. Basic single-loop scrubbing configuration with E.H.T.
modified to simulate tubular flow.
121
-------
5.0 5.5 6.0 6.5 7.0 7.S
Figure 2. Concentration of dissolved sulfite in RTF
scrubbing liquor as a function of pH.
122
-------
The rise in pH is accounted for by the reaction:
2H+ + CaC03 -»• Ca44" + C02 + H20 (1)
which consumes part of the H needed to keep the reaction cycle, that is
responsible for oxidation, going at constant pH:
HS03~ + i$02 ->• H+ + S04~~ (2)
H+ + CaS03(s) + Ca44" + HS03~ (3)
Ca44" + SO^" + 2H20 •+• CaS04-2H20 (4)
It was clear that additional H must be provided during the course of
reactions (2 - 3) if the pH was to be held constant.
STAGED SCRUBBING
The scrubber was modified to a two-stage configuration as shown in
Figure 3 to permit a constant.pH of 4.5 to be maintained in the oxidizer.
A "staged" scrubber is defined here as one in which each gas-contacting
stage has^its own hold tank and slurry recirculation loop. Ideally, there
would be no liquor carryover from the first to the second-stage loop by
gas entrainment, and this was the case in all experiments reported here.
The primary loop, in which most SO^ absorption occurs, consisted of a TCA
tower operating at 18 - 20 cm water AP, and was set up with multiple hold
tanks. The purge stream from this loop, which would normally go to the
clarifier or filter, is instead sent to the first stage loop where the
slurry is contacted with additional S02—at maximum partial pressure—to
provide the H+ needed for reaction (1). The low pH required for oxidation
and CaSO-(s) dissolution is thus maintained in the first stage with a
minimum sacrifice of S02 removal efficiency in the primary scrubber. Since
little additional S02 absorption is required in the first stage it was set
up as a spray tower with low pressure drop (1 - 5 cm water) and low L/G.
In full scale application the first stage would be a venturi; several
systems, comprising a venturi in combination with a second, more efficient
S0_ scrubber, are already in operation in the U.S.
123
-------
SO? 2.S3k«/lir 3830 ppm S02
HCI 73 j/hr
LIQUOR
TOTAL S AS S03 3.30 s/litet
S02 1.60
Ca 2.70
Cl 3.70
Ml 0.45
pH 4.4
SOLIDS
TOTAL SASS03 542m|/i
S02 330
CO 5
Ca 294
OXIDATION * 24 mol %
UTILIZATION » 92 mol %
SETT. RATE * 0.2 cm/min
8 liters/min
(L/6=8«al./mcf)
M
i
502 = 2970 ppm 1
|
»"
<»)
A
2.7
m/«~-
1
cmH20
S
1)7
26
I/I tor
pH3.8
\
-\
' \
1
£
340 mm
I
S02 = 588 ppm
(81%)
— •
•
.
»•
/
V
2.7
m/sec
nm
UUL.
NT
17
anH
no
1 1
in
300CD
iffli
"m
JLXJ
on
JU
pHS.O
\
\
1|21
i
S3 mers/min
(L/G = 64 gaL/mcf) ,
LIQUOR
TOTAL SASS03 1.90g/liter
S02
Ca
Cl
Mg
pH
0.52
1.40
2.30
0.33
5.2
SOLIDS 18%)
TOTAL S ASS03 519 mg/g
S02
C02
Cl
OXIDATION
343
35
3D5
= 17 mol'.
UTILIZATION = 85 mol %
LIMESTONE t FILTRATE
fate
0
\
\
n \2oiitets
^
-J
\
A
rieis
€
45min
2 min
2 min
2 min
Figure 3. Staged scrubber without forced oxidation.
124
-------
When forced oxidation is not attempted S07 builds up in the first
stage scrubbing liquor to the concentrations shown in a typical run in
this mode in Figure 3. In contrast, Figure 4 shows the result of forcing
oxidation in the first stage by installing an air sparger in the hold
tank. The accumulated S0_ is oxidized in the liquor, together with the
solid calcium sulfite that is fed from the primary loop. The latter is
redissolved, oxidized in the liquid phase, and recrystallized as gypsum.
Comparison of the liquor compositions shown in Figures 3 and 4 indicates
that the steady-state S02 concentration in the first stage is no longer
determined solely by pH and chloride when oxidation is forced: it is
determined instead by the difference between the rate of oxidation and
the rate of S0_ feed to the oxidizer.
The pH was controlled at 4.5 in the oxidizer by adjusting the rate
of limestone feed to the primary scrubbing loop. Since the utilization
was already 85% in that loop, the additional S02 absorption in the first
stage boosted the overall limestone utilization to values approaching
100%.
Figure 4 is the basic scrubber configuration used at RTF for forced
oxidation. A major difference between this configuration and that used
by MHI is the fact that oxidation was carried out as an integral part of
the first stage in the RTF tests. This has the advantage of providing a
slurry of essentially pure gypsum to control scaling in the first stage,
which is most susceptible to that problem. We were thus able to operate
at L/G = 1.3 l./m3 (10 gal./mcf) in the first stage without scaling when
set up according to Figure 4, but could not operate at this L/G when
oxidation was not forced in the first stage. In the latter case gypsum
scale formed in the upper parts of the tower where the feed slurry
"entered. Another advantage of our approach is the ability to control pH
at the optimum value throughout the course of the oxidation reactions.
This minimizes the stripping of SO- from the liquor in the oxidizer and
increases the oxidation efficiency to the extent that it can be conducted
at atmospheric pressure and without catalysts.
125
-------
SOZ 2.53 k|/h>
HCI 41 }flir ~
2700 ppm S02
SOz = 2470 ppm (9%)
9 liten/min
LiaUOB
TOTAL S AS S03 2.22 i/liter
SO 2 0.45
C02 0.20
Ca 1.85
Cl 3.3S
M| 0.93
pH 4.5
SATURATION RATIO = 1.2
TOTAL SASS03 545m|/|
SOZ S3
COz ' «
Ci 275
OXIDATION -MnralX
UTILIZATION = 99 moIX
TO FILTER
SETT. RATE = 1.6 em/mil.
SETT. DENSITY-0.70|/ml
(L/G = 10 pUmcf)
AIR 13 A k|/hr
/\
2.7
m/set
I
cmHzO
S02
1.1
pH4.2
SOIitm/kr
59 liters/min
(L/G = 64 jal./mcf)
LIQUOR
TOTAL SASS03 1.54|/liter
S02
C02
Ca
Cl
Mi
pH
0.20
0.22
1.40
2.64
0.71
5.8
SATURATION RATIO = 1.0
SOLIDS (8%)
TOTAL SASS03 S02m|/g
S02 299
COz 45
Cl 297
OXIDATION « 26 mol X
UTILIZATION • 85 mol X
8.67 k|/lir
LIMESTONE + FILTRATE
38.0% SOLIDS
2.2 mill
22mm
Figure 4. Staged scrubber with forced oxidation in the first stage
(0.91-meter slurry depth; air stoichiometry = 6.3).
126
-------
OXIDIZER EFFICIENCY
As indicated by the results shown in Figures 4 (no fly ash) and
5 (with fly ash), oxidation was successfully accomplished in the RTF
scrubber by air sparging the first stage hold tank. The tank used for
these initial tests was 71 cm in diameter and contained a 0.91-meter
2
depth of slurry. Air was admitted at 0.7 - 1.4 kg/cm pressure through
a 50-cm diameter ring containing twenty-two 1.6 mm holes. Experience
showed that the holes must point downward to prevent slurry from entering
the tube, where the solids settle and eventually fill it. A 6.35-mm
orifice was used to measure air flow. Water was added at 10 ml/min at
a point ahead of the orifice to inhibit scale growth in the small holes
of the sparge ring. Satisfactory results were obtained with sparge rings
made of 316 stainless, PVC, and Teflon. The Teflon ring appeared best
with respect to lack of corrosion and tendency to resist pluggage. The
oxidizer tank was vented td the scrubber inlet so the air feed pressure
represents sparge ring AP almost entirely.
Figures 4 and 5 show that the air feed rate required to complete
the oxidation of the slurry with the type of oxidizer described above
was 6-7 times the stoichiometric minimum, where air stoichiometry is
defined by:
, . g-atoms of oxygen fed to oxidizer
air stozchzometry = g_moles of sOj absorbed in scrubber
= (kg air fed/hr) 0.21 (64.1) 2
29 (kg S02 fed/hr) n
where n is the combined SO™ absorption efficiency of both stages.
Attempts to reduce the air stoichiometry below 6 resulted in a
buildup of HSO ~ in the first stage scrubbing liquor, indicating that
the rate of oxygen absorption had become the limiting factor—the rate
of oxidation in the liquid phase at pH 4.5 has been shown (Gladkii,
1974) to increase with increasing HS03~ concentration; therefore,
oxidation kinetics were not limiting. The HS03~ concentration rises
in the first stage because, in this case, it is a dependent variable
determined by the oxidation rate—and not vice versa. The steady-state
127
-------
St>2 2.S7 ki/hr
HCI 45 g/hr "
LIQUOR
TOTAL S AS S03 2.30 j/liter .
2910ppmS02
SO; = 2590 ppm
802 = 520 ppm
(82%)
S02
C02
d
Cl
Mg
PH
0.87
0.03
2.52
5.44
0.77
4.S
SATURATION RATIO = 1.2
SOLIDS
TOTAL S AS S03 346 mg/g
SO; 15
C02 S
Ca 179
OXIDATION = 95 mol %
UTILIZATION * 97 mol %
TO FILTER
SETT. RATE = 2.0 cm/min
SETT. DENSITY = 0.91 |/ml
TOTAL SASS03 '-82g/litef
SATURATION RATIO = 1.0
SOLIDS (15%)
TOTAL S AS SO 3 306 mg/g
S02 218
C02 30
Ca 186
OXIDATION =11 mol %
UTILIZATION = 82 mol %
11.4kg/hr
LIMESTONE + FLY ASH
1tmin
2.2 min
2.2 min
2.2 min
Figure 5. Forced oxidation with fly ash addition
(0.91-m slurry depth; air stoichiometry = 7.0)
128
-------
oxidation rate is limited, in turn, by the absorption of oxygen from
the air bubbles. The number of bubbles per unit volume of slurry
decreases in the oxidizer with the air feed rate and thus the air/liquor
mass-transfer surface limits the amount of oxygen that can be absorbed
during the time of contact between the bubbles and slurry.
The efficiency of oxygen mass transfer to a sulfite solution is
discussed by Urza and Jackson (1975) and their rationale was followed to
improve the oxidation efficiency of the RTF system. The absorption of
02 from air bubbles of a given size was shown to be proportional to liquor
depth (bubble residence time) for oxidizer heights to 16 meters. The
oxidation kinetics ir: solution influence the overall rate only via the
effect on the concentration of oxygen in the bulk liquid phase. From
this viewpoint the high HSO,, concentration at low pH enhances the
absorption of oxygen -by minimizing the concentration of dissolved 0- in
the liquor.
Accordingly, the stirred-tank oxidizer was replaced with a tower
(PVC pipe) having a static slurry depth of 3.2, and later, 5.5 meters.
As shown in Figures 6, 7 and 8 these changes permitted us to reduce the
air stoichiometry ultimately to 2.6. The air sparger used in the tower
had the same size and number of holes as that used in the stirred tank.
As a result of the increased bubble residence time afforded by greater
slurry depth in the oxidizer, oxygen transfer efficiency was improved
and more S07 was oxidized per kg of air injected. It is reasonable to
expect that stoichiometries lower than 2.6 can be achieved by further
increase of the oxidizer height. Also, as pointed out by Urza and Jackson,
one can expect a bonus in the form of a reduced energy requirement for
Q
air compression, per unit of O^ transferred.
As indicated by Figure 6 the residence time of the slurry in the
oxidizer (with respect to the rate of recirculation to the first stage
tower) was reduced to 7 min. without ill effect on either the oxidation
efficiency or scaling. The CaSO^l^O supersaturation of the liquor
increased, however, from l.lx at 16 min. to 1.3 - 1.Ax at 7 min. residence
time. The liquor saturation ratios shown in Figures 3-8, 11 were
determined at 25° C by the direct test method.
129
-------
S02 1.93 kg/b
HCI 1009/hr
15 liter
LIQUOR
TOTAL SASS03 1.76g/liter
S02 0.16
C> 4.82
Cl 9.10
Mg 0.79
pH 4.5
SATURATION RATIO = 1.4
SOLIDS (13.8%)
TOTAL S AS S03 355 mg/g
S02 6
C02 15
Ca 1M
OXIDATION =98 mol X
UTILIZATION « 96 mol %
TO FILTER
SETT. RATE » 2.6 an/out
SETT. DENSITY = 0.89 I/ml
(50% SOLIDS)
(L/G = 20
Al
5.2k
kg/cm2
13cm r—^
"IF
2840ppmS02
s/min
g»Umcfl
g/hr
^
3.2m
102 liters
51 °C
0
ff V
S02 = 2220 ppm
FOUR
80%
OPEN
GRIDS <
ASH
1.5 kg/hi
i^^
(22%)
"A
2.0
m/s«
JP
1.5
mH20
S02
0.72
I/liter
pH3.9
1 1 II t
451
-»•
ittn/hr
1
S02
2.0
rn/stc
5.6
pH5.0
\ '
i
Sin*
e?
= 650ppm
(77%)
51 littrs/min
(L/G = 71
TOTAL S
SO
CC
Ci
Cl
FOUR Mi
65% pr!
G°R,DS SATURA
SO
TOTAL S
S<
C
C
OXIOA
UTIU2
6.27kg
LIMESTONE +
38.8HSOl.II
\
rs VI 02 liters
gal./mcf) ,
LIQUOR
AS SO 3 1.56g/liter
2 0.13
12 0.12
4.31
6.79
1 0.61
5.3
TION RATIO = 1.3
LIDS (8.2%)
AS S03 423 mg/g
J2 235
D2 119
i 302
TION = 30 mol %
ATION = 70 mol %
/hr
FILTRATE
IS
\
^204 liters
\- £
Tana
3 mill
Figure 6. Oxidizer depth increased to 3.2 meters
(air stoichiometry = 3.2).
130
-------
SOZ 3.44 kg/h
HCI 36 g/hr
25 liters
LIQUOR |
TOTAL S AS SO 3 1.49g/lit«
S02 0.07
Ca 2.08
Cl 3.70
. Mg 0.42
pH 4.3
SATURATION RATIO = 1.1
SOLIDS (15.5%)
TOTAL SASS03 306mg/8
S02 12
C02 I*
Ca 167
OXIDATION = 95 mol %
UTILIZATION =92 mol %
TO FILTER
SSTT. RATE - 3.0 cm/min
SETT. DENSITY = 0.88 n/ml
(60% SOLIDS)
, 2880 ppm SOz
min
(L/G = 20 gil./mcf)
M
k*
25cm
H20
Al
7.851
osV
cir»2
L
-
R
g/hr
5.5m
400 liters
O
FL
S02 = 2460 ppm
1 ASH
3.4 kg/hr
~T>
1 5.8 mifl
A
(15%)
3.7
m/sec
S02
0.88
g/litet
pH3.6
87 liters/hr
i
» S02
I
A
3.7
mno
IXJUU
17
jQCCXX
nmn
CJULXJQ
pH4.6
- 770 ppm
73%)
79 lilers/min
(L/G = 64 9aUmd> i
LIC
TOTAL S AS
S02
C02
Ca
Cl
M8
pH
SAT UR ATI
SOLI
TOTAL S*
SO 2
COj
Ca
OXIDA1
UTILIZ/
LIMESTONE +
\ ..
\ '
Yl 80 lite
UOR
SQ3 1.61 g/liter
0.23
D.19
1.58
1.88
0.30
5.1
)N RATIO = 1.1
OS (8.6%)
iSSOa 463mg/g
277
79
304
ION = 25 mol %
\ftOH = 76 mol %
hr
FILTRATE
37.4% SOLIDS
V --
fs \180liters
— &£*
2.2 min 2.2 min
t_w
ters
€
2.2 min
Figure 7. Oxidizer depth increased to 5.5 meters
(air stoichiometry = 2.9).
131
-------
SO? 3.56 kj/hr 3030j>pmS02
HCI lOOj/hr
S02 = 2550 ppm
SO; - 590 ppm
(8«4)
LIQUOR
TOTAL SASS03 2.01 g/litet
(L,'G = 70gal./nuf)
LIQUOR
TOTAL SASSOl 1.489/liter
SATURATION RAT'O
SATURATION RATIO = 1.1
SOLID SPHERES
8.9 cm BED DEPTH
TOTAL SASS03 S36mj/g
SO; 5
CO; 10
Ca 277
TOTAL SASS03 494 mg/g
SO? 333
C02 64
C> 302
OXIDATION = 16 mol V
UTILIZATION = B2molS
!28k5/hf
LIMESTONE* FILTRATE
OXIDATION = 99 mol S
UTILIZATION . 97 mol %
SETT. RATE « 1.1 cm/mm
SETT DENSITY = 0.76 i/ml
(57% SOLIDS)
1S.I
3 min
2 mil*
4 min
Figure 8. No fly ash addition (5.5-m oxidizer depth;
air stoichiometry = 2.6).
132
-------
OXIDATION AND SETTLING RATE
An important objective of forced oxidation is to improve the
settling rate of the slurry and reduce the size of the clarifier
which, at Shawnee for example, is the largest single vessel in the
scrubber system. An improvement in the settling rate as a result
of oxidation was first reported by TVA (Kelso, 1971), an effect
attributed to enlargement of the crystal size. Data accumulated
at RTF during the course of this study (Figure 9) confirm TVA's
conclusion. Using Shawnee (Fredonia) limestone, the fully oxidized
slurry settled 10 times faster than sulfite slurries in the normal
range of oxidation. Our results also confirm that very high levels
of oxidation must be attained before the improvement becomes apparent.
SETTLED DENSITY AND SLUDGE PRODUCTION
The Aerospace sludge^disposal study (Rossoff, 1974) has pointed
out the importance of sludge density to its environmental acceptability.
This property is the major determinant of compaction strength, compres-
sibility, and permeability. One of the reasons for the interest in
forced oxidation is the possibility of producing a sludge that can be
dewatered by filtration to a moisture content near the optimum value
fqr compaction. Thus, the use of dry additives, that are required to
attain the optimum water content with sulfite sludge, might be avoided.
The load-bearing strength and drainability are both superior for a
compacted sulfate sludge, and a waste product in this form is more suit-
2
able than sulfite sludge for direct disposal as landfill.
Figure 10 shows the density of settled sludges obtained at RTF,
which clearly increased with oxidation to a value of 0.9 g dry solids
per ml of sludge volume (equivalent to 60% solids) when fully oxidized.
In actual practice, settled densities corresponding to 35 - 45% solids
are obtained with most sulfite sludges. The values of Figure 10 are
somewhat lower because the test was made under conditions that avoid
compression of the settled solids, and they were not mechanically
disturbed by a clarifier rake. It is not unlikely therefore that a
133
-------
2.0
1.5
oc
o
1.0
• WITH SHAWNEE FLYASH
O WITHOUT FLYASH
CO
0.5
20
40 60
PERCENT OXIDATION
80
100
Figure 9. Settling rate of RTF limestone scrubber slurries
as a function of oxidation (50° C).
134
-------
1.0
60
•a
s
1
M
i
V)
o
Ul
_l
IU
M
0.6
0.4
GO
50 §
o
CO
LU
O
oc
40
30
0.2
• WITHSHAWNEE FLYASH
O WITHOUT FLYASH
20
40 60
OXIDATION, percent
80
100
Figure 10. Settled (quiescent) density of RTF limestone
scrubber sludge, as a function of oxidation.
135
-------
large clarifier would yield a fully oxidized sludge of 65% solids as
expected.2 Vacuum filtration of the oxidized slurry yielded a filter
cake that averaged about 67% solids in our tests. These samples were
reslurried in acetone and dried at room temperature to ensure that no
water of hydration was lost. Thus, the objective of achieving 80%
solids by filtration does not appear attainable on the basis of RTF
2
results. This value is ideal for optimum compaction and, therefore,
is important to the question of direct disposal as a landfill.
It was also apparent from our tests that settled sludges containing
60% solids could only be obtained with the longer oxidizer residence
time (16 min., based on the slurry recirculation rate to the first stage),
which corresponds to about 5 hours residence time for the slurry in the
first stage. At shorter residence time (7 min.) we obtained only 50%
solids. This suggests that the additional crystal growth that can occur
at long residence time is important to achieving maximum settled density.
The combined effects of high limestone utilization and improved
dewatering properties of the sludge can be expected to have a significant
impact upon the rate of waste production by a power plant. The effects
are shown in Table 1 for various scrubber operating conditions, using
the present situation defined in the SOTSEP report as a basis for
comparison. It is clear from Table 1 that: a) about 24% reduction of
the amount of sludge can be obtained using standard gravity settling;
b) a greater reduction of sludge is achieved by the use of more efficient
dewatering equipment, such as vacuum filtration, than is possible by
forced oxidation alone—thus, forced oxidation is justified only if gravity
settling is used for dewatering.
FLY ASH
In general, settling rate is determined by the size of the smallest
particles in suspension. It is clear from Figure 9 that the settling
rate of the mixture of fly ash and CaS04/CaS03 was dominated by calcium
sulfite. The fly ash fed to the scrubber during these tests, fly ash
"A", had a settling rate of 5.5 cm/min at 50° C in the absence of
CaSO,/CaS03. Later tests with a different fly ash (fly ash "B", also
-------
Table 1, ANNUAL WASTE SLUDGE PRODUCTION BY A 1000 MW COAL-FIRED
POWER PLANT EQUIPPED WITH LIMESTONE FGD SCRUBBERS:
FLY ASH COLLECTED IN SCRUBBER
(SHORT TONS)
Dewatering Procedure
Scrubber
Operating Conditions
Coal Ash, Dry
CaSO "'-sH.O
CaC03
Totnl Dry Solids
Solids Moisture, %
Total Wet Sludge
Reduction of Sludge
Production
Settling
Utilization - 60%
Oxidation - 10%
338,000
322,000
48,000
185,000
893,000
50
1,790,000
Base Case
Settling
Utilization - 90%
Oxidation « 10%
338,000
322,000
48,000
30, BOO
739,000
50
1,478,000
17%
Oxidation/Settling
Utilization - 100%
Oxidation = 100%
338,000
0
477,000
0
815,000
40
1,482,000
24%
Filtration
Utilization - 902
Oxidation » 10%
338,000
322,000
48,000
30,800
739,000
38
1,192,000
33%
Oxidation/Filtration
Utilization - 100Z
Oxidation - 100Z
338,000
0
477,000
0
815,000
30b
1,164,000
•35Z
0-4
~-i
aSOTSEP Report1, p. 60
b,,.
Minimum moisture content obtainable by filtration in RTP tests
-------
from the Shawnee Test Facility) showed that the settling characteristics
of the scrubber slurry can be dominated by the fly ash instead of calcium
sulfite. Fly ash "B" settled at 1.5 cm/min in water at 50° C and in the
scrubber yielded a slurry that settled at a rate that was independent of
oxidation. Not only was the settling rate of this slurry determined by
the fly ash, but an adverse interaction was also apparent; e.g., a sample
of fully oxidized slurry taken from the scrubber when operating without
fly ash (and settling at a rate of 3 cm/min) settled at only 0.7 cm/min
when it was reslurried with fly ash "B" and compared at the same temper-
ature of 50° C. The settled density was also reduced by the presence of
fly ash "B", the fully oxidized sludge containing only about 50% solids.
In view of this experience it appears that the current industry trend
toward dry collection of the fly ash prior to the scrubber is well advised.
Although the pervasiveness of "B"-type fly ashes is not known, it is clear
that minimal benefit can be-expected from forced oxidation when it is
present in the scrubber slurry.
Dry fly ash collection can be considered as an alternative to forced
oxidation as a means of reducing the waste production and, if it is mixed
with the scrubber sludge, as a means of increasing the density. In view
of the results at RTF, it may be the only way to achieve the desired 80%
solids for optimum compaction. Table 2 summarizes the relative amounts
of waste produced under varying conditions when the fly ash is collected
ahead of the scrubber. Comparison of Tables 1 and 2 shows that the
reduction of total waste is greater for dry fly ash collection than is
possible when the ash is collected in the scrubber, and this is true
whether or not forced oxidation is employed. As indicated by Table 2,
however, blending the dry fly ash with the scrubber sludge can also yield
the desired 80% solids if the sludge is oxidized.
SO- ABSORPTION
As indicated in Figure 6 the scrubber was successfully operated
. with both stages set up as spray towers at a total pressure drop of
only 8 cm water and 77 - 80% S02 removal. When an L/G of 2.7 l./m
138
-------
Table 2. ANNUAL WASTE SLUUCE PRODUCTION BY A 1000 MW COAL-FIRED
POWER PLANT EQUIPPED WITH LIMESTONE FGD SCRUBBERS:
DRY FLY ASH COLLECTION
(SHORT TONS)
Dcwatcring Procedure
Scrubber
Onerntin;; Conditions
Coal Ash, Dry
CaSO •I:1I20
CaSO, -2H.O
!t 2
CaCO
Solids Moisture, %
Total Wet Sludjje
Total Waste
Reduction of Waste
Compared to Base Case
Settling
Utilization - 60%
Oxidation = 10%
338,000
322,000
48,000
185,000
50
1,110,000
1,448,000
19%
Settling
Utilization = 90%
Oxidation = 10% .
338,000
322,000
48,000
30,800
50
817,000
1,155,000
35%
Oxidation/Settling
Utilization » 100%
Oxidation • 100%
338,000
0
47 71, 000
0
40
795,000
1,133,000
37%
Oxidation/Filtration
Utilization - 1002
Oxidation « 100%
338,000
0
477,000
0
30
682,000
1,020,000
43%
Filtration
Utilization - 90Z
Oxidation * 10%
338,000
322,000
48,000
30,800
38
647,000
985,000
45Z
Final Sludge
Densi ly, 7, Solids
63
66
73
80
75
Assuming dry fly ash is blended with
wet slud)./.>.
-------
(20 gal./mcf) was maintained in the first stage, the S02 removal in that
stage averaged 20Z at a gas velocity of 2.0 m/sec. Thus, a utilization
of only 70-80% is required in the second stage in order to control pH
at the desired value, and this range of utilization is within the
capability of a single stirred tank. Figure 11 summarizes a test in
which one 9-min. hold tank was used in the second stage. Although a
reduction of the overall S02 removal efficiency (to 71%) occurred, the
test showed that complete oxidation could still be obtained.
Solid spheres were used in most of the tests at 3.7 m/sec gas
velocity to permit the L/G to be increased above the 6.7 l./m (50 gal./
mcf) that is normally the upper limit for hollow (5-g) spheres. The
solid (linear polyethylene) spheres were 2.5 cm diameter and formed three
beds, each bed 7.6 cm deep. This tower could be operated at L/G's as
high as 12 l./m without flooding at 3.7 m/sec. The improvement results
from the reduction of sphere'buoyancy, which causes the light spheres
to congregate against the upper retaining grid. An alternate solution
is to increase the grid spacing, which in our case was 91 cm. As shown
3
in Figures 7 and 8 the solid spheres gave 73% removal at L/G = 8.5 l./m ,
o
17 cm water AP and 81% removal at L/G = 9.4 l./m , 25 cm AP, when
operating at 3.7 m/sec.
CONCLUSIONS AND RECOMMENDATIONS
The high oxidation efficiency obtained by air sparging at atmospheric
pressure indicates that considerable improvement in performance of forced
oxidation systems can be realized if the oxidation step is carried out as
an integral part of the first stage. The improvement will permit the use
of simpler, less costly oxidizer designs, will reduce the power requirements
and will eliminate the need for catalysts. The air stoichiometry is
determined primarily by the height of the oxidizer and future tests should
be made with slurry depths to at least 18 meters. Stoichiometries below
2.6 can thus be expected, and a minimum power for air compression.
It is clear that many variations of the two-stage configuration are
workable, including two spray towers, multiple or single hold tanks, and
140
-------
SOz I.SSkjrt
HCI 36 9/hr
15 liters/
LIQUOR i
TOTAL SASS03 1.59g/liter
SOz °-34
Ca 2.69
Cl 4.97
M) 0.67
pH 4.6
SATURATION RATIO = 1.3
SOLIDS (15.1%)
TOTAL SASS03 347 mg/g
SOz 16
C02 14
Ca 184
OXIDATION =94 mol %
UTILIZATION = 94 mol %
TO FILTER
SETT. RATE = 2.0 an/mm
(L/C
fl
kg
10cm
H20
, = 20 ga
A
4.6 I
.7?V-
/cm*
If
, 2850 ppm S02
min
l./md)
R
5/hr
^
3.2m
102 liters
52°C
-o
FL\
SO
/ASH
1.63 kg/hr
"~Q
/
2 =
2350 p
(17.5%)
\
2.0
m/sec
1.5
cm H20
S02
0.83
g/liter
pH4.0
I
1
502
/^
2.0
6.6
cm H20
PH5.3
45liteis/ht
1
1
= 830 ppm
71%)
51 liters/min
(L/G = 70 jaL/mcf) j
LIQUOR
TOTAL SASS03 1.36g/litn
S02 0,19
COz 0.18
Ca 1.97
Cl 3.49
Mg 0.47
pH 5.6
SATURATION RATIO = 1.3
TOTAL SASS03 423 mg/g
S02 276
C02 118
Ca 313
OXIDATION =18 mol %
UTILIZATION = 70 mol %
t, OH Vn/hr
LIMESTONE + FILTRATE
l i
40.1% SOLIDS
460 liters
O
7 mm
9 min
Figure 11. Single stirred tank in primary scrubber loop.
141
-------
gas velocities ranging from 2.0 - 3.7 in/sec (6.7 - 12 fps). Our
results favor the following conditions: 3.7 m/sec gas velocity,
minimum AP in the first stage at L/G = 1.3 - 2.7 l./m , maximum
pressure drop (hold up) in the second stage, and multiple hold tanks
of 6 - 9 min total residence time in the second stage. The oxidizer
should have a residence time of 15 min to provide a settled sludge of
maximum density.
The benefits to be derived from oxidation, in terms of improvement
of the physical properties of the waste, will depend upon the properties
of the fly ash as well as the properties of the gypsum produced. In any
case our results do not indicate that a sludge filterable to more than
70% solids can be expected. Dry fly ash collection would, under these
circumstances, be the preferred method of minimizing waste production
as it would afford the only opportunity to produce a directly disposable
landfill.
The chemical properties of gypsum, vis-a-vis calcium sulfite as
9
the most desired waste product, is the principal issue upon which the
final decision must be made as to whether forced oxidation is to be
undertaken on a large scale in the U.S.
142
-------
REFERENCES
1. Princiotta, F.T., "Sulfur Oxide Throwaway Sludge Evaluation Panel
(SOTSEP): Final Report, Volume II" EPA-650/2-?5-OlO-b (NTIS No.
PB 2U2-619/AS), April 1975.
2. Rossoff, J., et al., "Disposal of By-Products from Non-Regenerable
Flue Gas Desulfurization Systems," Proceedings: Symposium on Flue
Gas Desulfurization-Atlanta Nov. 197k, Volume I" EPA-650/2-7h-126-a
(NTIS No. PB 2^2-572/AS), pp. 399-kkl, December 1971;.
3. Uno, T., et al., "The Pilot Scale R&D and Prototype Plant of MHI
Lime-Gypsum Process," Proceedings of Second International Lime/
Limestone Wet-Scrubbing Symposium, pp. 83U-8U9, November 1971.
lu Gladkii, A.V., et al., "State Scientific Research Institute of
Industrial and Sanitary Gas Cleaning (Moscow),11 Report for Protocol
Point A-1, Development of Lime/Limestone Scrubbing for Stack Gas
Desulfurization, US/USSR, Sulfur Oxides Technology Sub-Group, 197k.
5. Berkpwitz, J.B., et al., "Scale Control in Limestone Wet Scrubbing
Systems," EPA-650/2-75-031 (NTIS No. FB 2U3-309/AS), pp. 52-56,
April 1975.
6. Wen, C.Y., et al., Ibid., p. 66, April 1975.
7. Borgwardt, R.H., "Increasing Limestone Utilization in FGD Scrubbers,11
paper presented at 68th Annual Meeting, AIChE. Los Angeles Nov. 1975.
8. Urza, I.J., and Jackson, M.L., "Pressure Aeration in a 55-ft Bubble
Column," Ind. Eng. Chem., Process Des. Dev., 1_5 pp. 106-113,
April 19751
9. Kelso, T.M., et all, "Limestone Wet-Scrubbing Pilot Plant at Colbert
Steam Plant," Process Engineering Branch Report, Tennessee Valley
Authority, April 1971.
143
-------
RESULTS OF MIST ELIMINATION AND ALKALI UTILIZATION TESTING AT
THE EPA ALKALI SCRUBBING TEST FACILITY
M. Epstein, H. N. Head, S. C. Wang, and D. A. Burbank
Bechtel Corporation
50 Beale Street
San Francisco, California 94105
ABSTRACT
Testing with a single stage, 3-pass, open-vane, 316 stainless
steel chevron mist eliminator in both the venturi/spray tower and
TCA systems with lime and limestone has shown that the accumulation
of soft solids on the mist eliminator is a strong function of alkali
utilization. For high alkali utilization (greater than about 85 percent),
the mist eliminator was kept free of solids deposits by use of inter-
mittent fresh water wash on both the top and bottomside. For alkali
utilization less than about 85 percent, intermittent top and bottomside
wash with fresh water did not limit solids accumulation. However, for
these conditions, a continuous bottom wash with diluted clarified
liquor used in combination with an intermittent topside fresh water
wash was shown to limit soft solids buildup to less than 10 percent
restriction within the mist eliminator.
In the venturi/spray tower system with lime additive, reliable
long-term operation of the chevron mist eliminator with intermittent
top and bottomside fresh water wash was demonstrated during constant
load operation at the maximum attainable spray tower gas velocity of
9.4 ft/sec and during variable load (cycling gas rate) operation. The
lime utilization for these tests was about 90 percent. In the TCA
system with limestone additive, reliable long-t^rm operation of the
chevron mist eliminator with intermittent top and bottomside fresh
water wash was demonstrated at 12.5 ft/sec gas velocity at high alkali
utilization (greater than about 90 percent). For both the lime and
limestone testing, only about half the makeup water available during
closed liquor loop operation was needed to wash the mist eliminators.
Utilization with lime in the venturi/spray tower system was
normally greater than 85 percent for scrubber inlet liquor pH less
than 9.0. Utilization with limestone in both scrubber inlet liquor
pH of 6.0 to about 95 percent at a scrubber inlet pH of 5.2. Operation
at reduced scrubber liquor pH, however, inherently causes a reduction
in SO removal efficiency. For the venturi/spray tower system with
a single effluent hold tank, limestone utilization was not affected
145
-------
by a change in residence time from 20 to 12 minutes. Limestone
utilization declined, however, at 6 minutes residence time for scrubber
inlet liquor pH greater than about 5.6. For the TCA system at 12
minutes total residence time and at scrubber inlet liquor pH greater
than about 5.0, higher limestone utilization was achieved with three
hold tanks in series than with a single hold tank.
146
-------
CONTENTS
Section Page
1 INTRODUCTION 149
2 ADVANCED TEST PROGRAM OBJECTIVES
AND SCHEDULE 157
3 LIME TESTING WITH A CHEVRON MIST
ELIMINATOR IN THE VENTURI/SPRAY
TOWER SYSTEM 161
3. 1 Testing at Constant Gas Velocity 161
3.2 Variable Load Testing 168
3. 3 Conclusions 170
4 LIMESTONE TESTING WITH A WASH TRAY AND
CHEVRON MIST ELIMINATOR IN SERIES IN
THE TCA SYSTEM 172
4. 1 Testing at 8.6 ft/sec Superficial
Gas Velocity 172
4.2 Testing at 10 and 12 ft/sec Superficial
Gas Velocity 175
4. 3 Conclusions 176
5 LIMESTONE UTILIZATION TESTING IN THE
VENTURI/SPRAY TOWER AND TCA SYSTEMS 177
5. 1 Utilization Testing in the Venturi/Spray
Tower System with Variable Residence Time 178
5. 2 Utilization Testing in the TCA System with
Three Hold Tanks in Series 184
5. 3 Conclusions 193
6 EFFECT OF ALKALI UTILIZATION
ON MIST ELIMINATOR OPERABILITY 196
6. 1 Summary of Mist Eliminator Operability
During Lime and Limestone Reliability
Testing 196
147
-------
Section Page
6. 2 Mist Eliminator Operability During
Limestone Utilization Testing 198
6. 3 Conclusions 203
7 REFERENCES 204
148
-------
Section 1
INTRODUCTION
In June 1968, the EPA initiated a program to test prototype lime and
limestone wet-scrubbing systems for removing sulfur dioxide and par-
ticulates from coal-fired boiler flue gases. The program -was con-
ducted in a test facility integrated into the flue gas ductwork of boiler
No. 10 at the Tennessee Valley Authority (TVA) Shawnee Power Station,
Paducah, Kentucky. Bechtel Corporation of San Francisco was the
major contractor and test director, and TVA was the constructor and
facility operator.
The results of testing at the facility during the original program,
nrfaich lasted from March 1972 to May 1974, are presented in References
1 and 2. The most significant reliability problem encountered during
the testing was associated with scaling and/or plugging of mist elim-
ination surfaces.
In June 1974, the EPA, through its Office of Research and Development
**
and Industrial Environmental Research Laboratory initiated a three-
year advanced test program at the Shawnee facility. Bechtel Corporation
•is continuing as the major contractor and test director, and TVA as
The National Air Pollution Control Administration prior to 1970,
The Control Systems Laboratory prior to 1975.
149
-------
the constructor and facility operator. The major goals established
for the advanced program are: (1) to continue long-term testing with
emphasis on demonstrating reliable operation of the mist elimination
systems at increased gas velocity, (2) to investigate advanced process
and equipment design variations for improving system reliability and
process economics, and (3) to perform long-term (2 to 5 month)
reliability testing on promising process and equipment design variations.
Two parallel scrubbing systems are being operated during the advanced
program:
• A venturi followed by a spray tower
• A Turbulent Contact Absorber (TCA)
Each system has its own slurry handling facilities and is capable of
treating approximately 30,000 acfrn of flue gas from the TVA Shawnee
coal-fired boiler No. 10. This gas rate is equivalent to approximately
10 Mw of power plant generating capacity. Boiler No. 10 normally
burns a high-sulfur bituminous coal which produces SO_ concentrations
of 1500 to 4500 ppm and inlet particulate loadings of 2 to 4 grains/scf
in the flue gas.
Figures 1-1 and 1-2 (drawn with major dimensions to scale) show the
two scrubber syste/ns along with the mist elimination systems currently
used for de-entraining slurry in the exit gas streams. The cross-
sectional area of the spray tower is 50 ft^ in both the scrubbing section
and mist elimination section. The cross-sectional area of the TCA is
ty O '
32 ft^ in the scrubbing section and 49 & in the mist elimination section.
150
-------
CHEVRON MIST
ELIMINATOR
SPRAY TOWER
GAS IN
I
INLET SLURRY wt^X
THROATA
ADJUSTABLE PLUG
YENTURI SCRUBBER
A>!
! PLUG/ U
GAS OUT
«
WWW
A A TV
A A A
MIST ELIMINATOR
WASH WATER
MIST ELIMINATOR
WASH LIQUOR
A A A
A A TV
r
t.
INLET SLURRY
5'
APPROX.SCALE
EFFLUENT SLURRY
Figure 1-1. Schematic of Venturi Scrubber and Spray Tower
151
-------
MIST ELIMINATOR
WASH WATER
CHEVRON MIST
ELIMINATOR
RETAINING IAR-GRIDS
6AS IN
GAS OUT
T U
A 7\ A
o—„- -
o
>
000
00 0°0
O °0
|2.0£L
MIST ELIMINATOR
WASH LIQUOR
•INLET SLURRY
MOIILE PACKING SPHERES
i 1
APPROX. SCALE
EFFLUENT SLURRY
Figure 1-2. Schematic of Three-Bed TCA Scrubber
152
-------
The TCA utilizes a fluidized bed of 1 1/2 inch diameter, light weight
(5.0-6. 5 gram) spheres which are free to move between retaining grids.
The mist elimination systems shown in Figures 1-1 and 1-2 each
consists of a 3-pass, open-vane chevron mist eliminator with provision
for underside and topside washing. During the early portion of the
advanced test program, the TCA was also tested with a mist elimination
system consisting of a wash tray (Koch Flexitray) in series with a
6-pass, closed-vane chevron mist eliminator, both with underside wash.
A typical system configuration used during venturi/spray tower lime
reliability testing is shown in Figure 1-3. A typical system con-
figuration used during TCA limestone utilization testing is shown in
Figure 1-4. In the TCA configuration shown, three effluent hold tanks
in series were used to increase limestone utilization.
This paper presents the results of mist elimination and limestone
utilization testing at the Shawnee facility from June 1974 through
January 1976. During this period the venturi/spray tower system was
operated with both lime and limestone and the TCA system with
limestone. All of the testing was performed under closed liquor loop
conditions. Due to the relatively high inlet gas particulate loading, the
slurry solids contained about 40 to 50 wt % fly ash.
For all of the tests during the reporting period, the sulfate (gypsum)
saturation of the scrubber inlet liquor was maintained below about
140 percent and there was no significant accumulation of gypsum scale
on scrubber internals. Previous testing has shown that scrubber in-
ternals can be kept relatively free of gypsum scale if the sulfate
153
-------
tn
THMB
1
PROCESS
WATER
HOLD
TANK
STACK
Discharge
SETTLING POND
O Gas Composition
® Particulate Composition & Loading
0 Slurry or Solids Composition
.. _ Gas Stream
_ Liquor Stream
Figure 1-3. Typical Process Flow Diagram for Venturi/Spray Tower System
-------
en
O Gas Composition
® Participate Composition & Loading
® Slurry or Solids Composition
_ _ Gas Stream
— Liquor Stream
SFJTLING POND
Figure 1-4. Typical Process Flow Diagram for TCA System
-------
saturation of the scrubber liquor is kept below about 135 percent (see
Section 7. 3 in Reference 1). The effects of the process variables on
scrubber liquor sulfate saturation observed during the advanced test
program will be presented in a future paper.
156
-------
Section 2
ADVANCED TEST PROGRAM OBJECTIVES AND SCHEDULE
The Shawnee Advanced Test Program is scheduled to run from June
1974 through June 1977. The objectives of the advanced program are:
• To continue long-term testing with emphasis on demonstrating
reliable operation of the mist elimination systems.
• To investigate advanced process and equipment design varia-
tions for improving system reliability and economics. For
example, testing will be conducted to investigate the practical
upper limit of the gas velocity (i. e. , minimum scrubber
size) at which the scrubber mist elimination systems can be
reliably operated. Also, tests will be conducted to evaluate
system performance under conditions of minimum energy
consumption for the desired levels of SO2 and particulate removal.
• To evaluate process variations for substantially increasing
limestone utilization and reducing sludge production. Tests
will be conducted with scrubber effluent passing through
three stirred tanks in series to approach a "plug-flow" con-
dition and at reduced scrubber liquor pH to increase lime-
stone utilization (see Reference 3).
• To evaluate scrubber operabiltiLy during variable load (e.g. ,
cycling gas rate) operation.
• To perform long-term (2 to 5 month) reliability testing on
advanced process and equipment design variations.
157
-------
To investigate methods of improving waste solids separation.
This may include testing a multiple-plate thickener, using
coagulants, attempting to relate sludge characteristics to
operating conditions, and making operational improvements
on the centrifuge and filter.
To study schemes for oxidizing sludge in order to improve
solids settling characteristics and to reduce the chemical
oxygen demand (COD) of the sludge. Oxidation tests will be
conducted using a two-stage scrubber system developed by
Borgwardt (Reference 4).
To evaluate the effectiveness of three commercially offered
sludge fixation processes and of untreated sludge disposal.
Fixed sludges (Chemfix, Dravo, and IUCS) and untreated lime
and limestone sludges are being continuously monitored in
ponds at the Shawnee site. Aerospace Corporation is the major
contractor and test director for this effort.
To evaluate system performance and reliability without fly
ash in the flue gas. Tests will be conducted with flue gas
taken downstream of the Shawnee boiler No. 10 electrostatic
precipitator, i. e., with less than 0. 1 grain/scf of particulate
in the inlet flue gas.
To determine the practical upper limits of SO2 removal
efficiency. Tests will be conducted to determine the practical
upper limit of SOo removal by increasing the scrubber slurry
pH, increasing the slurry rate, increasing the scrubber gas
pressure drop, and adding magnesium ion (MgO) to the slurry.
To evaluate the TCA performance with lime and the venturi/
spray tower performance with limestone.
To characterize stack gas emissions including outlet par-
ticulate mass loading and size distribution, slurry entrain-
ment, and total sulfate emission.
To evaluate, under the direction of TVA, corrosion and wear
of alternative plant equipment components and materials.
To develop a computer program, in conjunction with TYA, for
the design and cost comparison of full-scale lime and limestone
systems.
158
-------
The current test program schedule, based on the defined objectives,
is presented in Figure 2-1. As can be seen in the figure, as of
January 1976, limestone tests were in. progress on both the venturi/
spray tower and TCA systems to: (1) demonstrate the reliability of
the mist elimination systems and (2) determine the effects of process
variables on limestone utilization.
159
-------
LIMtSTONt ADVANCED THTiNfl WTH 1CA |Y|TtM
MflT tLM
tOXlDI AODttlOM ritTtNQ
ALKALI UTILIZATION UITINQ WITHOUT MfO ADDITION
rACTofllAl TtlTINQ WITHOUT MfO ADDITION
FACTORIAL TESTINQ WITH MfO ADDITION
VARIABLE 10 AD Tt IT I WO
FLV ASH FNtE TtlTINO
Rlt(A»KITV OfMOMTHATKMf
MINIMIZt ENtflav UTILIZATION TUT (WO
UME ADVAftCI
RELIABILITY TESTING
FACTORIAL TE*T1NQ
UME ADVANCED IWTIWO WJTH Vf NTUfll/JWUV T0W|« SVJTtM
MIST ILIMINATOfl TESTING
MAGNESIUM OXIDE ADDITION TESTING
VARIABLE LOAD TESTING
FACTORIAL TESTING'*'
FLY ASH FRtE TESTING
RELIABILITY Of MONSTNATION flUN
MINIMUM ENFftOV UTILIZATION TCSTINO
LIMESTONE
MtlT ftlMINAT»R TISTtNC
MADNESILtM DXIDE ADDITION TESTING
ALKALI UTILIZATION TESTINO WITH AND WITHOUT MfO ADDITION
FACTORIAL TfST ING WITHOUT MfO ADDITION
FACTORIAL TESTING WITH M« ADDITION
ESTINO
VARIABLE LOAD TESTING
tESTINQ COMMON TO BOTH TRAINS
SLUDGE CHARACTERIZATION
STACK OAB I MISSION CHARACTf RIZATION
6 TESTING
7 GENERAL PUBLICATION AEPOHT DRAF T SUBMITTAL DATES
111 INCLUDES "MAXIMIZING SOj REMOVAL EFFICIENCY TESTING! MAY INCLUDI ADDITION OF MfOI
NOTE OASMfD LINES HEfRESI NT SEQUENCES WHICH MAT BE CONTINUED DURING THE HniOO INDICATED
Figure 2-1. Current Schedule for Shawnee Advanced Test Program
-------
Section 3
LIME TESTING WITH A CHEVRON MIST ELIMINATOR
IN THE VENTURI/SPRAY TOWER SYSTEM
This section summarizes the results of mist elimination reliability
testing from June 1974 through mid-October 1975 with lime on the
venturi/spray tower system. During this period, a 3-pass, open-
vane, 316 stainless steel chevron mist eliminator was used in the
spray tower.
3. 1 TESTING AT CONSTANT GAS VELOCITY
During the period of advanced testing from June 1974 through
March 1975, several mist eliminator washing configurations were
evaluated under the following typical test conditions:
Spray tower gas velocity 6.7 ft/sec
Venturi liquid-to-gas ratio 30 gal/mcf
Spray tower liquid-to-gas ratio 60 gal/mcf
Weight percent solids recirculated 8
Effluent residence time 12-24 minutes
Scrubber inlet slurry pH (controlled) 8
Weight percent solids in discharge cake 43-60
Inlet gas SO_ concentration 1150-4250 ppm
*
•In this report, all gas velocities and'liquid-to-gas ratios are at scrub-
ber operating conditions, i. e. , saturated gas at scrubber temperature.
With flue gas operation, the scrubber temperature is approximately
125°F. The gas velocities are all superficial velocities.
161
-------
For these test conditions, SO_ removal was 75 to 95 percent, lime
;';
utilization was approximately 90 percent, the scrubber outlet liquor
pH was approximately 5. 0, the sulfate (gypsum) saturation of the
scrubber inlet liquor was 105 to 140 percent, the dissolved solids
in the scrubber liquor was 6000 to 12000 ppm, and the total pressure
drop was about 12. 5 inches HO, including 9 inches HO across the
venturi.
.JL...JU
Initially (Runs 604-1A through 608-1A") the mist eliminator was
washed from the bottomside only. With this wash configuration, soft
solids accumulation did not occur on the mist eliminator, but there
was a continual problem of gypsum scale formation on the top vanes.
The runs showed that this scale formation rate on the top mist elim-
inator vanes (25 to 50 mils/week) was relatively unaffected by the
wash cycle, wash rate, or quality (sulfate saturation) of the wash
liquor. Run 608-1A showed that intermittent washing with high
pressure (45 psig) raw water at a rate of 3 gpm/ft2 for 9 minutes
every 4 hours gave results that were at least as good as results
with continuous washing with relatively low pressure raw water at
0. 3 gpm/ft2 (Run 606-1A). This is significant in that intermittent
washing may be required in closed liquor loop operation due to restric-
tions in the allowable raw water makeup to the scrubber system.
*Percent alkali utilization is defined as 100 x moles SO2 absorbed
per mole of Ca added. The reciprocal of the fraction alkali utiliza-
tion is stoichiometric ratio, defined as moles Ca added per mole
of SO2 absorbed.
**Detailed operating conditions for Runs 604- 1A through 624- 1A can be
found in References 1 and 2.
162
-------
Starting in late September 1974, runs were made (Runs 609-1A and
2
610-1A) in which the entire mist eliminator underside and a 14 ft
»t-
area on the topside were washed at high pressure (45 psig) with
makeup water at a rate of 2. 7 gpm/ft for the underside and 1. 0
gpm/ft for the topside in an intermittent cycle of about 8 minutes
every 4 hours. After 530 hours of operation, the washed area of
the top mist eliminator vanes was essentially clean (less than 1 mil
of solids accumulation) compared with an average of 70 mils scale
buildup on the rest of the topside vane surfaces. The mist eliminator
bottom vanes were covered with smooth white scale about 10 mils
thick.
In the final run at 6. 7 ft/sec spray tower gas velocity (Run 623-1A),
begun in early March 1975, a wash system was installed in the spray
tower to provide sequential and intermittent washing of the topside
vanes of the chevron mist eliminator. The topside wash was ac-
complished by operating 6 nozzles in sequence. Every 80 minutes,
one nozzle was actuated for 4 minutes at a rate of 0. 5 gpm/ft at
2
13 psig. The underside wash rate was 3 gpm/ft at 45 psig for 3. 5
minutes every 4 hours.
*
Only a small section of the topside was washed because of a concern
that entrainment from the top spray might overload the reheater and
possibly allow moisture to reach the fan. In later tests, the entire
topside was washed by resorting to sequential washing. Another
possible solution would have been to use a second mist eliminator
to intercept entrainment from the topside sprays.
163
-------
After 162 hours of operation, an inspection showed that the combina-
tion of sequential topside wash and intermittent underside wash was
successful in preventing scale and solids accumulation on the mist
eliminator. Only scattered dust a few mils thick was formed on some
of the vanes.
A new run (Run 624-1A) was started on March 19, 1975, without system
cleaning, to test the mist elimination system at a higher spray tower
gas velocity of 8. 0 ft/sec. For this run, the intermittent wash cycle
for the underside of the mist eliminator was increased from 3. 5
to 4. 3 minutes every 4 hours because of the higher makeup water
rate available at the higher gas velocity. The wash rate and cycle
for the topside were not changed from the previous run. Run 624-1A
was terminated on April 23, 1975, after 823 hours of operation due
to a scheduled 5.-week maintenance outage on boiler No. 10.
A total of 4 inspections were made for this run, including one at the
end of the run. The mist eliminator was found to be free of any
restriction during each of these inspections. Vane surfaces washed
earlier in the wash cycle usually held 2 mils of scattered white dust,
while recently washed vane surfaces were entirely clean. Operating
data for the initial 480 hours of Run 624-1A are presented in Figure 3-1.
Following the boiler outage, a run was begun on June 20, 1975, (Run
625-1A) to continue testing the mist eliminator at 8 ft/sec spray tower
gas velocity with a reduced underside wash rate of 1. 5 gpm/ft for 4. 3
minutes every 4 hours. The wash rate and cycle for the topside were
the same as for the previous two runs. Only about one-half of the makeup
164
-------
; BEGIN RUN 62* 1;
3.500
3.°°°
is ""
= " 2.000
3,500
3.000
2,500
2.000
1.500
I 322 I 3/23 I 3;24 I 3/25 I 3/26 I 1/27 I 3/Z8 I 3/29 I 3/30 I 3/31 I
CALENDAR DAV
jj "
: :, u
1
1 0
Hi 30
1
£ W
2 10
o
150
1 ,0.
s «
.
_ . ^ ^
"*"" /^ ^ ^^\ ^ ..*__-- S^-~--~--S * -^*-
* " ^ * ^
— f
'" ' J>v r-'"''"
* -A- *^^" / ' / *
* " ^ ^ ' '"* '/^v/N/ '
-
-
*" * -A. /*" *
'— " , _--- — -^-/
• TOTAL DISSOLVED SOiLDS NOTE SPECIES WHOSE _,
9,000
i B.OOO
1 1 '•M0
z e 6,000
3 O 5,000
§ 9
Q ^ 4.000
0 Z J.OOO
5 2,000
1.000
0 O CALCIUM (C»**l CONCENTRATIONS ARE LESS
• • U SULFATE ISO/) THAN 600 pom ARE NOT
A CHLORIDE Cl h PLOTTED jr
•**• », . •*. -
•* •«•••*
A A*
* * ** »» » A A » A * A*
- - °° ° I °'° 'y' ' ^ oo * » o o o o o o o ° > SS
_t cf C Z Dn_
i i i i i i i " i i i i
1.3
12
1.1
1.0
30
20
10
0
ISO
100
so
10.000
9,000
8,000
7.000
6.000
5,000
4.000
3.000
2,000
1.000
0 40 80 1 20 1«0 20O 240 280 320 360 400 440 «80
TEST TIMS Houn
CALENDAR DAY
Gas Rite; 30.000 acfm ^ 330 °F
Liquor Rate lo Veniuri ~- 600 gpm
Liquor Rate lo Spray Tower ; 1200 gpm
Venturi L/G = 25 gal/mcf
Spray Tower L/G = 50 gal/met
Sprav Towei Gas Vetocilv = S.OIt/iec
No. of Spray Headers = 4
EHT ftendence Time = 17 min
Peicent Soiids Recirculated = 7-10 wrt %
Venturi Pressure Drop - 9 in HjO
Total Pressure Drop, Excluding Mist Elim. - 12.5-12.6 m
ScrubbQr Inlet Liquor Temperature - 120-126 °F
Liquid Conductivity - 6,100-10,000 u, mhos/cm
Discharge (Clarifier and Filter) Solids
Concentration = 4S-&4 wt %
Lime Addition to Scrubber Downcomer
Figure 3-1. Operating Data for Venturi/Spray Tower Run 624-1A
165
-------
water available during closed liquor loop operation -was needed to wash
the mist eliminator. After 319 operating hours, the chevron mist
eliminator was essentially clean (less than 2 percent restricted with
dust).
The spray tower gas velocity was increased to the maximum achieve-
able value of 9.4 ft/sec in the next run (Run 626-1A) which started on
July 9, 1975, with no cleaning of the mist eliminator. The inter-
mittent wash cycle for the underside of the mist eliminator was
increased to 6 minutes every 4 hours because of the higher makeup
•water rate available at the higher gas velocity. The topside wash rate
and cycle were unchanged. As in the previous run, only about one-
half of the makeup water available during closed liquor loop operation
was needed to wash the mist eliminator. An inspection after 569 hours
«
of operation showed that the chevron mist eliminator had remained
essentially clean during the test (less than 2 percent restricted).
Operating data for the initial 480 hours of Run 626-1A are presented
in Figure 3-2.
The recirculation slurry solids concentration was increased from
8 wt % to 15 wt % in Run 627-1A which began on August 5, 1975.
The spray tower gas velocity was maintained at 9.4 ft/sec. During
this run, the inlet gas SO_ concentration dropped to a low value
^ .o-
'i*
of about 1500 ppm from August 6 through August 8. Asa consequence
during that period, the outlet scrubber liquor pH rose to about 6. 2
(the scrubber inlet liqiior pH was controlled at 8. 0). An inspection
*
Low sulfur Montana coal was being burned in boiler No. 10.
166
-------
If*
cii
10.000
9,000
2 5 5-°°°
82
O -J 4.000
r
U CULFATE I004-|
* CHLOR1OE ICI " •
AHE L£S3 IM.-kN 500,1.1
.\HE NOT PLOTTED
. •
.•• .
6.000
S.OOO
3000
7.000
7 ?« I l 26 I ' Z6 I 1 ?7 I 7 28 I
GaiRiti* 35.000 acfm * 330 °F
Spray ToMtr G« Velocity " 9.4 fl/HC
Liquor Rat» to Venturi - 600 gpm
Liquor Rati to Spray Tower • 1,400 gpm
V«ntutlL/G-21sd/mc(
Sony Tomr IVG - 50 pl/mct
No. of Spny Hndtn • 4
EHTRMlH.net Tirni - 12 mln
Percent Solid! Recirculsterl - 8-9 wt %
Venturi Pressure Drop • 9 in. H^O
Ton) Prenure Drop. Excluding Milt Film. - 14.2-15 in.
ScrublMi Inlet Uqupr Tempmturt • 130-133 °F
Liauid Conductivity • 11,000-17,000 u mhoi/cm
Ditcharte IdirlHer ind Film) Solid!
Concentration - 57-60 wt %
Ume Addition to Scrubbw Oowwcomw
Figure 3-2. Operating Data for Venturi/Spray Tower Run 626-1A
167
-------
on August 8 showed that the higher level of liquor pH within the spray
tower resulted in calcium sulfite scale formation within the tower
-i-
and on the lower vanes of the chevron mist eliminator. An inspection
at the end of the run, after 187 hours of operation, showed that the
scale had diminished somewhat during the latter half of the test, and
that the mist eliminator was 2 to 3 percent restricted with sulfite
scale and dust. At the conclusion of Run 627-1A, the mist eliminator
had been in operation for 1075 hours without cleaning, including 319
i
hours at 8. 0 ft/sec and 756 hours at 9.4 ft/sec gas velocity.
3. 2 VARIABLE LOAD TESTING
Run 628-1A was begun on August 16 to test the operability and con-
trollability of the venturi/spray tower system under cycling gas load.
The gas flow rate for this run (17, 000 to 35, 000 acfm) was varied with
the actual Unit No. 10 boiler load (about 60 to 160 Mw), resulting in
spray tower gas velocities between 4. 5 and 9. 4 ft/sec. Constant liquor
rates were used in the venturi and spray tower. The pressure drop
across the venturi was held constant at 9 inches HO for all gas rates
L+
by varying the plug position. The same mist eliminator wash scheme
as in Runs 626-1A and 627-1A was used during the test. Operating
data for the initial 480 hours of Run 628-1A are presented in Figure 3-3.
As can be seen in the figure, the inlet gas SO_ concentration varied
between 1500 and-4200 ppm during the test.
This problem of sulfite scale formation at high outlet scrubber
liquor pH was corrected in subsequent testing by controlling the
outlet scrubber liquor pH to 5. 0_+0. 5, with the constraint that the
scrubber inlet liquor pH remain equal to or less than 8. 0.
168
-------
: MGINRONU*!*
M -i
V
sC^
GH Rate - 17.000-3B.OOO aclm S> 330 °F
Spray Tower G« Velocity « 4.6-9.4 ll/stc
Liquor Rate to Venturi - 600 gpm
Liquor Rate to Spray Tow«r - 1.400. 1.600 [after 8/25)
Venturi L/C - 21-««al/"
-------
An inspection after 717 hours of operation showed that the mist elim-
inator was essentially clean (only 2 percent restricted with dust). No
problems were experienced in controlling the system during the test.
Rurl 628-1B was a continuation of 628-1A, except that the venturi plug
position was fixed to give a maximum venturi pressure drop of 9
inches H O at the maximum gas rate of 35, 000 acfm. Operating
Ct
data for this test are presented in Figure 3-4. An inspection after
426 hours of operation (for a. total of 1143 hours for Runs 628-1A and
628-IB) showed that the mist eliminator -was essentially unchanged
(2 percent restricted with dust)'.
3. 3 CONCLUSIONS
The long-term (6-12 month) operability of the 3-pass, open-vane chevron
mist eliminator in the venturi/spray tower system with lime additive
has been demonstrated for intermittent high pressure underside washing
and sequential low pressure topside washing with raw water. The
mist eliminator remained essentially free of solids deposit for 756
hours of operation at a spray tower gas velocity of 9. 4 ft/sec and with
8 to 15 wt % solids in the recirculating slurry. The mist eliminator
also remained free of solids deposit during variable load (cycling
gas rate) tests lasting a total of 1143 hours. For the lime testing,
only about one-half of the makeup water available during closed liquor
loop operation was needed to wash the mist eliminator. The lime
utilization for these tests was about 90 percent.
170
-------
T~
11}
I'l
i!
si
Gas Rate * 19,000 • 35.000 acfm @ 330 °F
Spray Tower Gas Velocity ~ 5.1-9.4 ft/sec
Liquor Rate to Ventun ~ 600 gpm
liquor Rate to Spray Tower - 1.600 gpm
Ventun L/G - 21-39 gal/md
SpravTowet L/G - 57 105 gal/met
No. ot Spray Headers - 4
EHT Residence Time -12mm
Percent Solids ftecircuiated - 8.69.7 wl %
Ventun Pressure Drop 4.0-9.0 m. HjO
Total Preiiure Drop, Excluding Misi Elim. - 6.5-14.6 in.
Scrubber Inlei Liquor Temperature = 129'T31 °F
Liquid Conductivity " 8,000-11,500 u. rnhos/cm
Oisctiarge (Clanfier and Filter) Solids
Concentration - 52-56 wt %
Lime Addvion lo Scrubber Downcomer
Figure 3-4. Operating Data from Venturi/Spray Tower Variable
Load Run 628-1B
171
-------
Section 4
LIMESTONE TESTING WITH A WASH TRAY AND
CHEVRON MIST ELIMINATOR IN SERIES IN THE TCA SYSTEM
This section summarizes the results of mist eliminator reliability
testing from September 1974 through March 1975 with limestone in
the TCA system. During this period the mist elimination system
consisted of a wash tray (Koch Flexitray) in series with a 6-pass,
closed-vane, 316 stainless steel chevron mist eliminator.
4. 1 TESTING AT 8. 6 FT/SEC SUPERFICIAL GAS VELOCITY
TCA limestone Run 535-2A was begun on September 12, 1974, and
continued through December 4, 1974, for a total of 1835 operating
*
hours. The TCA internals consisted of 3 beds (4 grids), with 5
inches per bed of 1 1/2 inch diameter, 6 gram TPR (thermoplastic
rubber) spheres. The major test conditions at the start of the
run were:
Gas velocity 8. 6 ft/sec
Liquid-to-gas ratio 73 gal/mcf
Percent solids recirculated 15
Effluent residence time 12 min
Percent SO- removal (controlled) 84
Inlet gas SO? concentration 2000-4000 ppm
Weight percent solids in discharge sludge 35-42
i
TCA Run 535-2A is described in detail in Reference 2.
172
-------
For these test conditions, the scrubber inlet liquor pH varied between
5. 7 and 6. 0, limestone utilization was approximately 65 percent
(stoichiometric ratio of 1. 54 moles Ca added/mole SO7 absorbed),
the sulfate (gypsum) saturation of the scrubber inlet liquor averaged
110 percent, the dissolved solids concentration in the scrubber
liquor was 4000 to 8000 ppm, and the total pressure drop (including
the mist elimination system) was 6 to 7 inches HO. Operating data
for the initial 480 hours of testing are presented in Figure 4-1.
The clarified liquor return flow rate was maintained at a minimum of
15 gpm for Koch tray feed and mist eliminator wash to prevent high
sulfate supersaturation in the wash tray effluent liquor. The under-
side of the mist eliminator was washed continuously with 15 gpm (0. 31
2
gpm/ft ) diluted clarified liquor (about 9 gpm makeup water plus about
6 gpm clarified liquor). The tray was fed with the 15 gpm mist elim-
inator wash plus the remaining /^ 9 gpm (minimum) of clarified liquor
2
(0. 50 gpm/ft ). The underside of the tray was sparged with 125 psig
steam for one minute each hour.
Throughout the first half of Run 535-2A, there was a continual problem
of solids accumulation on the scrubber walls between the steam sparger
and the main slurry spray header. This problem -was resolved by
installing four wall spray nozzles utilizing the wash tray effluent liquor.
It was observed that the wall sprays were also effective in washing
the underside of the tray -where it was covered by the sprays.
After 1835 operating hours for Run 535-2A, an inspection showed that
the top surface of the wash tray was entirely clean and that the chevron
173
-------
40 H 120 ISO im 240 !BO
21 I 9?? I 9 ?3 I 9>?4 ! 9 2S I 9'76 I 9'2J I 9/28 I 9'M I 9/30 I 10'1 I
1 7
l] "
\'s '"
^ M
? | i.«
1 3
S t K
S f
H J M
Si to
g * "
13000
-
i '•
til 000
cc 10.000
; 9.000
^ 8000
£
£ 7.000
1
s '•M0
? 5.000
S
3 4000
S
2 3,000
S 2000
1.000
0
^ ' — ~ " / \
' \ " "\ (\ /
- /' "• V /' \ T'^' w i \ /^\ /v •' •
^ \ '' ^' ^- ' \ /
* V \ / "^ "*'i«-/^
"" -• • \ .'
' ' \ / -* ~* A
A . \y
- \^ ? ' ^ \ ^.^ \ ^ * ^ ^ ^"* .j » y v \ "
x " %
,
• TOTAL DISSOLVED SOLIDS
«"• CALCIUM (C.'M
ri SULFATE IS<>3 '
A CHLORIDE (Cl I
NOTE SPECIES WHOSE
CONCENTRATIONS ARE LESS
THAN WO ppm ARf. NOT
PLOTTED
*
9
• • • 9 f
-A * * A A •, A &
-8@3*o£ fiSn^o^g* §D f 300 °F
Liquor Rale- I200gpm
L/G = 73 gal/mcl
Gas Velocity = 8.6 It/sec
EHT (Sealedl Residence Time • 12 min (9/12-9/27),
15 mm (alter 9/271
T^ree Stages. 5 in spheres/stage
Petcent Solids Recitculated - 12-15 wt %
Total Pressure Drop. Excluding Mist Elim.
and Koch Tray = 4.0-4.6 in H^O
Scrubber Inlet LiQuor Temperature =120-126 °F
Liquid Conductivity • 4.800-10,000 u rr.hos/cn
Discharge IClsrifierl Solids
Concentration - 3642 wl %
Figure 4-1. Operating Data for TCA Run 535-2A
174
-------
mist eliminator was less than 5 percent restricted by solids (mostly
fly ash). The underside of the wash tray had no solids accumulations
greater than 1/4 inch thick. These deposits did not interfere with tray
operation. Also, the scrubber walls between the tray and slurry spray
header were clean.
Run 535-2A was arbitrarily terminated on December 4, 1974, and Run
535-2B begun, when low sulfur Montana coal was unexpectedly introduced
into Shawnee boiler No. 10. The wide-swings in inlet gas SO2 concentra-
tion (1200 to 4900 ppm in 8 hours) caused system upsets which resulted
in some scale formation on the wash tray and solids accumulation
in the chevron mist eliminator. During Run 535-2B, the steam sparger
*
was removed and the four wall spray nozzles were replaced by a single
tray under wash nozzle, which also provided rinsing of the walls below
tiie tray. This modification (see Figure 3-2 in Reference 2) proved to
be effective in keeping both the underside of the wash tray and the walls
below it clean.
4.2 TESTING AT 10 AND 12 FT/SEC SUPERFICIAL GAS VELOCITY
At scrubber gas velocities of 10 ft/sec or higher, plugging of the chevron
mist eliminator by solids (mostly fly ash) became a problem. At 10
ft/sec scrubber gas velocity (6. 5 ft/sec in the mist eliminator and wash
tray areas), the overall restriction of the mist eliminator increased
gradually to about 8 percent during the first 500 hours of operation and
In a full scale gas scrubbing unit, steam cleaning of the underside of
a wash tray -would be an economic burden.
175
-------
appeared to level out at 6 to 8 percent after 562 operating hours (Run
538-2A). At 12 ft/sec scrubber gas velocity (7.8 ft/sec in the mist
eliminator and •wash tray areas), the mist eliminator was 11 percent
restricted by solids within only 215 hours of operation (Run 539-2A).
4. 3 CONCLUSIONS
The long-term operability of the TCA mist elimination system was
demonstrated in an 1835-hour run in limestone service at 8.6 ft/sec
superficial gas velocity and 15 wt % solids in the recirculating slurry.
The mist elimination system consisted of a Koch Flexitray in series
with a 6-pass, closed-vane, chevron mist eliminator, both with under-
side wash. Long-term operability was not demonstrated at increased
gas velocity. At 10 ft/sec the chevron mist eliminator was 8 percent
restricted with soft solids after 562 operating hours, while at 12 ft/sec
the chevron mist eliminator was 11 percent restricted in only 215 hours.
176
-------
Section 5
LIMESTONE UTILIZATION TESTING IN THE
VENTURI/SPRAY TOWER AND TCA SYSTEMS
The results of limestone utilization testing from October 1975
through January 1976 on both the venturi/spray tower and the TCA
systems are presented in this section. These tests were made as a
result of a TVA economic study which showed that a potential existed
for significantly improving the economics of limestone scrubbing
by improving the utilization of the limestone feed. Improved lime-
stone utilization not only results in a decrease in limestone feed re-
quirements but also a corresponding decrease in waste sludge production.
Tests were conducted primarily to determine the effect of scrubber
inlet liquor pH, effluent residence time, and hold tank design on
limestone utilization. These variables were correlated with stoichio-
metric ratio (the reciprocal of fraction alkali utilization). For the
venturi/spray tower system a single backmix effluent hold tank was
used. For the TCA system both a single backmix hold tank and three
backmix hold tanks in series (to simulate a plug flow reactor) were
used.
For each combination of residence time and hold tank design, tests
were conducted to cover a range of values of scrubber inlet liquor
pH. Normally, the systems were run for about 4 to 5 days at a
177
-------
specified level of pH. During testing, the stoichiometric ratios were
determined every 4 hours from solids analyses of the scrubber re-
circulation sVurry.
The limestone used during the advanced program was "Fredonia Valley
White", purchased from Fredonia Quarries, Fredonia, Kentucky. The
limestone analysis was: 96 wt % CaCOy 1 wt % MgCO , and 3 wt %
inerts. The limestone size distribution was: 90 wt % less than 325
mesh, 87 wt % less than 30 microns, 82 wt % less than 22 microns,
and 52 wt % less than 6 microns.
Results of analyses presented here should be considered preliminary
as the data reduction is still in progress. For example, no attempt
has been made in this report to adjust the pH for chloride concentration
in the slurry liquor, which ranged from 1500 to 6500 ppm during the
testing.
Both the venturi/spray tower system and the TCA system were operated
with a single stage, 3-pass, open-vane, 316 stainless steel mist elim-
inator with top and bottom wash. Details of the mist eliminator systems
and the effect of alkali utilization on mist eliminator operation are
discussed in Section 6.
5. 1 UTILIZATION TESTING IN THE VENTURI/SPRAY TOWER
SYSTEM WITH VARIABLE RESIDENCE TIME
In the venturi/spray tower system, limestone tests were made with
a single backmix effluent hold tank to determine the effect of scrubber
178
-------
inlet liquor pH and residence time on stoichiometric ratio. Runs
were made at 20, 12, and 6 minutes residence time. Major test con-
ditions maintained constant during this period were:
Spray tower gas velocity 9.4 ft/sec
Venturi liquid-to-gas ratio 21 gal/mcf
Spray tower liquid-to-gas ratio 50-57 gal/mcf
Venturi pressure drop 9 in. HO
Percent solids recirculated 15
Data showing the relationship between stoichiometric ratio and scrub-
ber inlet liquor pH for the venturi/spray tower system are plotted in
Figures 5-1, 5-2, and 5-3 for 20, 12, and 6 minutes residence time,
respectively. As would be expected, scatter in the data was greatest
at the shortest residence time, where pH recovery time is limited and
a small variation in tank level can result in a significant change in
residence time.
Sight average curves drawn through the data in Figures 5-1 through
5-3 are cross plotted in Figure 5-4, which shows the effect of effluent
hold tank residence time and scrubber inlet liquor pH on stoichio-
metric ratio. Comparing 20 minutes residence time versus 12 in
Figure 5-4, little effect of residence time on stoichiometric ratio
tan be seen. However, between 12 minutes and 6 minutes, a signifi-
cant increase in stoichiometric ratio with decrease in residence time
can be seen for scrubber inlet liquor pH's greater than or equal to
about 5. 6.
All but 2 tests were at 50 gal/mcf.
179
-------
1.80
1.70-
1.60 •
V50!
8
1.40- •
2 1-30 +
u
c
Ul
I 1.20
u
i
1.10
1.00
4.80
8
VENTURI/SPRAY TOWER SYSTEM
SINGLE HOLD TANK
20 MINUTES RESIDENCE TIME
5.00
__l , , ,_
5.20 5.40 5.60 5.80
SCRUBBER INLET LIQUOR pH
6.00
6.20
Figure 5-1. Stoichiometric latio versus Scrubber Inlet Liquor pH for
a Single Hold Tank at 20 minutes Residence Time
180
-------
1.80
1.70 -
4.80
5.00
5.20 5.40 5.60 5.80
SCRUBBER INLET LIQUOR pH
6.00
6.20
Figure 5-2. Stoichiometric Ratio versus Scrubber Inlet Liquor pH for a
Single Hold Tank at 12 minutes Residence Time
181
-------
1.80
1.70- •
1.60- •
CM
8
1.50- •
1.40- •
!'•*>+
ec
u
ui
O 1.20-
O
i
1.10 •
1.00 - •
VENTUR I/SPRAY TOWER SYSTEM
SINGLE HOLD TANK
6 MINUTES RESIDENCE TIME
°00°
—t—
5.20
4.80
5.00
5.40 5.60
SCRUBBER INLET LIQUOR pH
5.80
6.00
ft 20
Figure 5-3. Stoichiometric Ratio versus Scrubber Inlet Liquor pH for a
Single Hold Tank at 6 minutes Residence Time
182
-------
1.5
IOMETRIC RATIO.
•d/mola SO 2 absorbed
-* ^
CO *
M
I | 1.2
1.1
1.0 -
i r- 1 1 1 i 1 1
VENTUR I/SPRAY TOWER SYSTEM
SINGLE HOLD TANK
^
\
o — «. i
~~""*-c> ~—o '
• ' ' 1 1 1 1 I
6 8 10 12 14 16 18 20
EFFLUENT HOLD TANK RESIDENCE TIME, MINUTES
Figure 5-4. The Effect of Effluent Residence Time and Scrubber Inlet
Liquor pH on Stoichiometric Ratio
183
-------
Figure 5-5 shows the general relationship between SO-, removal and
stoichiometric ratio for an inlet gas SO_ concentration range between
LJ
2500 and 3500 ppm. Data points have been plotted for hold tank residence
times of 20, 12, and 6 minutes and a sight average line through all
the data has been drawn. Referring back to Figure 5-4, it was shown
that, for a given stoichiometric ratio greater than about 1.25, the
scrubber inlet liquor pH is lower at 6 minutes residence time than at
12 or 20 minutes. Therefore, a corresponding reduction in SO_ removal
at 6 minutes residence time was expected. However, within the scatter
of the data, such a decrease could not be discerned in Figure 5-5.
Figure 5-6 shows the relationship between SO- removal and scrubber
inlet liquor pH for an inlet gas SO- concentration range from 2500
to 3500 ppm and a 12 minute hold tank residence time. Averages from
Figure 5-6 and from similar plots for higher and lower inlet gas SO
concentration ranges at 12 minutes residence time are drawn in Figure
5-7, which shows the effect of inlet gas SO concentration and scrubber
inlet liquor pH on SO removal. As expected, an increase in inlet
C*
SO concentration results in a decrease in SO removal at constant
L* £*
inlet pH (see Equlation 14-7 in Reference 1).
5. 2 UTILIZATION TESTING IN THE TCA SYSTEM WITH
THREE HOLD TANKS IN SERIES
Kinetic theory shows that for a continuous system where the reaction
order is greater than zero, raw materials are more completely con-
verted in a plug flow reactor than in a backmixed reactor of the same
residence time. This concept for improving utilization was success-
184
-------
100
95-
90 -
85 •-
80 --
cc
(SI
2
Ul
O
EC
ui
a.
70--
65 -
60 -•
55 -•
50
n
9
VENTURI/SPRAY TOWER SYSTEM
INLET GAS SO2 CONCENTRATION BETWEEN 2500 & 3500 ppm
SYMBOL
SINGLE HOLD TANK
RESIDENCE TIME
20 minutes
12 minutes
6 minutes
B
1.00 1.20 1.40 1.60 1.80 2.00
STOICHIOMETRIC RATIO, moles Ca added/mole SO2 absorbed
2.20
Figure 5-5. The Effect of Stoichiometric Ratio and Effluent Residence
Time on Percent SO, Removal with a Single Hold Tank
185
-------
100
95 •
90 •
85
80
cc
8'5
o
oc
70
65
60
55
SO
VENTUR I/SPRAY TOWER SYSTEM
SINGLE HOLD TANK
12 MINUTES RESIDENCE TIME
INLET GAS SO2 CONCENTRATION
BETWEEN 2500 & 3500 ppm
O
O
o o
•+•
4-
•4-
4-
4.80 5.00 5.20 5.40 5.60 5.80
SCRUBBER INLET LIQUOR pH
6.00
6.20
Figure 5-6. The Effect of Scrubber Inlet Liquor pH on Percent SO2
Removal in the Venturi/Spray Tower System
186
-------
100
95--
90--
86--
80--
UJ
-------
fully tested by Borgwardt (Reference 3) in a 0. 1 Mw limestone scrub-
ber with both a plug flow reactor and with 3 stirred tanks in series
to approximate plug flow. The concept has now been tested with lime-
stone in the TCA system at the Shawnee Test Facility (see Figure 1-4)
Major conditions for these tests were:
TCA superficial gas velocity 12. 5 ft/sec
Liquid-to-gas ratio 50 gal/mcf
Number of TCA beds ^ 3
Static sphere height per bed 5 in.
Percent solids recirulated 15
Hold tank configurations tested were:
Single hold tank at 12 minutes residence time.
Three hold tanks in series at 12 minutes total residence
time (4. 3, 2. 2, and 5. 5 minutes).
Three hold tanks in series at 9 minutes total residence time
(3. 8, 1. 9, and 3. 3 minutes).
Results of the utilization testing for these 3 configurations are plotted
as stoichiometric ratio versus scrubber inlet liquor pH in Figures
5-8, 5-9, and 5-10, respectively. Sight drawn averages from Figures
5-8 through 5-10 are compared in Figure 5-11. For the runs with
3 tanks in series, there was no significant difference between 9 minutes
and 12 minutes residence time. However, there was a distinct difference
Initially 6-gram hollow TPR (thermoplastic rubber) spheres were
used. These were subsequently replaced with 6. 5-gram nitrile-PVC
solid foam spheres.
188
-------
1.80
1.70 -•
1.60 -
i 1.50 - •
1.40 - -
W "*"
2
]
01.20 • •
i
i
1.10 • •
•1.00 • •
TCA SYSTEM
SINGLE HOLD TANK
12 MINUTES RESIDENCE TIME
o_ 8
o p° 1 o
00 0
o
o
-f-
-H
+
•+-
4.80
5.00
5.20 5.40 5.60 5.80
SCRUBBER INLET LIQUOR pH
6.00
6.20
Figure 5-8. Stoichiometric Ratio versus Scrubber Inlet Liquor pH for
a Single Hold Tank at 12 minutes Residence Time
189
-------
1.80
1.70 •
1.60
cw
2
"5
1.50
1-40
21.30
tc
u
cc
Ul
g 1.20
X
o
o
fc
1.10
1.00
4.80
TCA SYSTEM
THREE HOLD TANKS IN SERIES
12 MINUTES RESIDENCE TIME
—I—
5.00
+
_| 1 .—|
5.20 5.40 5.60 5.80
SCRUBBER INLET LIQUOR pH
6.00
6.20
Figure 5-9. Stoichiometric Ratio versus Scrubber Inlet Liquor PH for
Three Hold Tanks in Series at 12 minutes Residence Time
190
-------
1.80
1.70-
1.60-
N
8
1.50-
1.40-
1.20-
1.10 • •
1.00-
TCA SYSTEM
THREE HOLD TANKS IN SERIES
9 MINUTES RESIDENCE TIME
o
4.80
5.00
5.20 5.40 5.60
SCRUBBER INLET LIQUOR pH
5.80
6.00
6.20
Figure 5-10. Stoichiometric Ratio versus Scrubber Inlet Liquor pH for
Three Hold Tanks in Series at 9 minutes Residence Time
191
-------
1.80
1.70 •
1.60
1 1-50
N
S
1.40
<3
8
1.30
-------
between operation with a single tank and with 3 tanks in series
at pH levels greater than about 5. 0. For example, at a scrubber
inlet liquor pH of 5. 6, stoichiometric ratio averaged 1. 19 with a single
hold tank as compared with 1.11 with 3 tanks in series, a 6 percent
improvement in limestone utilization. At higher pH, the improvement
was greater. For instance at 5. 8 pH the improvement in utilization
was 14 percent.
The improvement in utilization -with 3 tanks in series can also be
seen in Figure 5-12 where SO_ removal is plotted versus stoichio-
Cj
metric ratio for an inlet gas SO concentration range of 2500 to 3500
C*
ppm. For example, at 85 percent SO_ removal and 12 minutes total
residence time, the stoichiometric ratio -with 3 tanks in series averaged
about 15 percent lower than with a single tank.
Data for 9 minutes residence time with 3 tanks in series were not
included in Figure 5-12 because the TPR spheres in the TCA were
replaced with nitrile-PVC solid foam spheres during the 9 minute
testing. Unfortunately, the effect of 9 minutes residence time on S®i
removal was confounded with the effect of the new bed of spheres on
SO,, removal.
5.3 CONCLUSIONS
Limestone utilization in the venturi/spray tower and TCA systems
normally varied from about 60 percent at a scrubber inlet liquor pH
of 6. 0 to about 95 percent at a scrubber inlet pH of 5. 2. Operation at
reduced scrubber liquor pH, however, inherently causes a reduction
in SO- removal efficiency.
193
-------
100
95 -
90 •
85 •
3 TANKS
80
i
Ul
E
(M
8
IU
U
c
70
65
60
55
1TANK
TCA SYSTEM
12 MINUTES RESIDENCE TIME
INLET GAS S02 CONCENTRATION
BETWEEN 2500 & 3500 ppm
SYMBOL
D
HOLD TANKS
50
1.0 1.2 1.4 1.6 1-8
STOICHIOMETRIC RATIO, moles Ca added/mole SO2ab*orb*d
2.0
Figure 5-12. The Effect of Stoichiometric Ratio and Hold
Tank Configuration on Percent SO2 Removal
2.2
194
-------
For the venturi/spray tower system with a single effluent hold tank,
limestone utilization was not affected by a change in residence time
from 20 to 12 minutes over the measured range of scrubber inlet liquor
pH. Limestone utilization declined, however, at 6 minutes residence
time for scrubber inlet liquor pH !s greater than about 5. 6.
For the TCA system at 12 minutes total effluent residence time and
at scrubber inlet liquor pH's greater than about 5.0, higher limestone
utilization is achieved with 3 hold tanks in series than with a single
hold tank.
195
-------
Section 6
EFFECT OF ALKALI UTILIZATION
ON MIST ELIMINATOR OPERABILITY
Data showing the effect of alkali utilization on mist eliminator oper-
ability are presented in this section.
6. 1 SUMMARY OF MIST ELIMINATOR OPERABILITY
DURING LIME AND LIMESTONE RELIABILITY TESTING
During lime testing on the venturi/spray tower system, reliable
operation was achieved for the 3-pass, open-vane, 316 stainless steel
chevron mist eliminator (see Section 3). These tests were conducted
at spray tower gas velocities up to 9.4 ft/sec (maximum attainable)
and recirculated slurry solids contents up to 15 percent. The lime
utilization for these tests was about 90 percent. Both the topside
and the bottomside of the mist eliminator were washed intermittently
with fresh water. Only about one-half of the makeup water available dur-
ing closed liquor loop operation was required for mist eliminator wash.
During limestone testing on the TCA system with the wash tray and
6-pass, closed-vane, chevron mist eliminator in series (see Section 4),
reliable operation of the mist elimination system was achieved only
at 8. 6 ft/sec scrubber gas velocity (5. 6 ft/sec superficial gas velocity
in the wash tray and mist eliminator areas). The limestone utilization
ranged from 60 to 75 percent.
196
-------
In view of the success of the spray tower mist eliminator operation
with lime, a. new mist eliminator of similar design (single stage,
316 stainless steel, 3-pass, open-vane, chevron type) was installed
in the TCA in July 1975 for testing with limestone slurry. However,
poor operability with this mist eliminator was experienced with inter-
mittent topside and bottomside wash with raw water. For example,
at a scrubber gas velocity of 9.4 ft/sec (6. 1 ft/sec superficial gas
velocity in the mist eliminator area), the chevron mist eliminator
became about 50 percent restricted with soft solids after only 60 hours
of operation (Run 554-2A). The limestone utilization was about 63
percent at the scrubber inlet liquor pH of about 6. 0. This failure
in the TCA, as compared with the successful operation in the spray
tower with lime using a similar mist eliminator design and washing
scheme, was initially attributed to the differences in the physical design
and the pattern of mist generation within the two scrubbers.
During subsequent testing of the TCA system at scrubber inlet pH
from 5. 7 to 6. 0 (limestone utilizations from 58 to 74 percent), stable
conditions with less than 10 percent solids restricted were achieved
for the mist elimination system by continuous washing of the under-
side of the chevron with diluted clarified liquor. For these low
utilizations, only the mist eliminator blades continuously contacted by
•wash water could be maintained free of solids restriction. It was
necessary, therefore, to eliminate shadowing by mist eliminator
supports and to avoid maldistribution of gas caused by observation
ports and corrosion coupon racks. In Run 559-2A, the chevron mist
eliminator was only 7 percent restricted with soft solids after 384
operating hours using continuous underwash with diluted clarfied liquor
2
at 0.4 gpm/ft and intermittent topside wash with raw water.
197
-------
6. 2 MIST ELIMINATOR OPERABILITY DURING LIMESTONE
UTILIZATION TESTING
In October 1975, limestone utilization testing was started on both the
TCA and venturi/spray tower systems (see Section 5), concurrent
with the on-going limestone mist eliminator testing in the TCA. Tables
6-1 and 6-2 summarize the results of mist eliminator operability during
these tests through January 1976. These tables list the average scrub-
ber inlet pH, average limestone utilization and stoichiometric ratio,
mist eliminator wash scheme, and percent of the mist eliminator
passage restricted by solids deposit at the end of each run. All of
the available makeup water was used for tests with continuous mist
eliminator bottom wash, while only about one-half of the available
makeup water was used for tests with intermittent bottom wash.
During venturi/spray tower limestone Runs 701-1A and 702-1A (see
Table 6-1) with intermittent topside and bottomside raw water wash, the
mist eliminator was heavily restricted by soft solids within 2 to 3 days.
This was unexpected since earlier operation wifh lime slurry under
identical operating conditions was successful. The limestone utiliza-
tion for these two runs was only 69 percent, as compared to 90 percent
normally obtained with lime operation. Subsequently, during Run
703-1A, the limestone utilization was increased to 93 percent by drop-
ping the average scrubber inlet pH to 5. 2. The mist eliminator was
found to be essentially clean (1 percent restricted) after 319 operating
hours with intermittent underside and topside raw water wash. The
average SO_ removal for Run 703-1A was only 58 percent, as com-
pared with 87 percent for the previous tests.
198
-------
Table 6-1
SUMMARY OF VENTURI/SPRAY TOWER LIMESTONE
UTILIZATION AND MIST ELIMINATOR TESTS
All rung made with a 316 stainless steel, 3-pasa, open-vane,
chevron mist eliminator. Run conditions: 1 5 wt % solids in
recirculated slurry, 9.4 ft/sec spray tower superficial gas
velocity, 21 gal/mcf liquid-to-gas ratio in venturi, 50-57
gal/mcf liquid-to-gas ratio in spray tower.
Run
No.
701-1A
702-1A
703-1A
704- 1A
705- 1A
706- 1A
707-1A
708-1A
709-1A
710-1A
711-1A
711-1B
712-1A
712-1B
713-1A
Single Tank
Res. Time,
min
20
20
20
20
20
12
12
12
12
12
6
6
6
6
6
A vg.
Scrubber
Inlet pH
5.9
5.8
5.2
5.8
5. 7
5. 2
5. 7
5. 6
5.9
6. 0
5. 7
5. 6
5-8
Depletion
5.2
A vg.
Stoich.
Ratio
1.45
1.45
1.07
1.45
1.25
1. 06
1. 30
1. 20
1. 35
1. 50
1. 30
1.40
1. 50
-
1. 10
Avg. Percent
Limestone
Utilization
69
69
93
69
80
94
77
83
74
67
77
71
67
-
91
Avg. Percent
SO2 Removal(g)
88
87
58
86
84
58
83
77
83
91
81
82
87
-
69
Mist Eliminator Run
Wash Scheme Hours
Top | Bottom
Intermittent Intermittent 73
\
60
'
319
66
136
180
Intermittent10' Intermittent!^' 118
1
Intermittent 138
Continuous 134
i
234
144
71
119
( 18
52
Hours Since
Cleaning Mist
Eliminator
73
60
319
385
136
180
298
436
134
368
512
583
702
720
772
Percent Mist
Eliminator
Restriction
50-60
45-50
1
45-50
17-20
1
10-15
15-20
<1
5-7
_
5-7
10-15
-
10
(a) Intermittent, sequential top wash with makeup water at 0. 53 gpm/sqft for 4 min/8 hr/section.
(b)
(c)
Intermittent, full face bottom wash with makeup water at 1. 5 gpm/sqft for 6 min/4 hr.
Intermittent, sequential top wash with makeup water at 0. 53 gpm/sqft for 3 min/hr /section.
(d) Intermittent, full face bottom wash with makeup water at 1. 5 gpm/sqft for 4 min/hr.
(e) Continuous, full face bottom wash with diluted clarified liquor at 0.4 gprr./sqft.
(f)
A limestone depletion run is conducted without limestone addition during SC>2 absorption.
The scrubber inlet liquor pH is allowed to drop from about 5. 9 to 4. 8.
(g) SC>2 removals are for 2500 to 3500 ppm inlet gas SC>2 concentration.
-------
Table 6-2
SUMMARY OF TCA LIMESTONE UTILIZATION AND
MIST ELIMINATOR TESTS
All rum rmde with a 316 itainlen steel, 3-p«n, open-vane, chevron mitt eliminator.
Run condition!: 15 wt % solids in reclrculated ilurry, 12. 5 ft/sec TCA superficial gat
velocity, 50 gal/mcf liquld-to-gai ratio. TCA configuration: 3 beds with 5 tnche« per
bed of 6-gram TPR spheres for Runs 562-2A through 569-2A and 6. 5-gram nitrile foam
spheres for Rani 569-2B through 582-2A.
Run Nc
No. Tan
562-2A
562-2B
563-2A
564-2A
Total
ki Res. Time,
mln
12
12
12
12
565-2A 3 12
566-2A 3 12
567-2A 3 12
568-2A 3 12
569-2A 3 9
569-2B
J 9
570-2A 3 9
571-2A 3 9
571-2B 3 9
572-2A
! 9
573-2A 3 12
573-2B 3 12
574-2A 3 9
575-2A
12
576-2A 3 12
576-2B
( 12
577-2A 3 12
578-2A 3 12
579-2A 3 9
580-2A 3 9
581-2A
582-2A
12
12
Avg. A
Scrubber St
Inlet pH R
5.9 1
5. 7 I
5.9 1
5.2 1
5.25 1
5. 9 1
6.0 1
5. 5 1
5. 5 1
5. 5 1
5. 75 1
5.8 1
Depletion'
5.25 1
5. 5 1
Depiction
Depletion
5.55 1
5. 7 1
Depletion
5.8 1
Depletion
5.25 1
Depletion
5. 5 1
Depletion
vg. Avg. Percent
olch. Limestone
atlo Utilization
.6 63
.4 71
. 7 59
.06 94
. 05 95
.25 80
.4 71
.06 94
. 08 93
.08 93
. 18 85
.25 80
.09 92
. 08 93
-
-
. 15 87
. 17 85
-
.25 80
-
.05 95
-
.15 87
-
Avg. Percent Milt Eliminator Run
SO, Removal'*' Wa.h Scheme Hours
Top ] Bottom
83 Intermittent'*1 Continuou«'b> 495
81
86
58
58
83
84
63
69
64
72
73
-
57
134
182
113
109
166
138
Intermittent ' 162
66 No wash
-
-
69
73
/. i
83 Intermittent'"
-
61
-
80
-
66
97
144
96
11
154
45
28
12
47
112
3
159
9
62
12
164
18
Houra Since
Cleaning Mist
Eliminator
155
332
495
629
811
924
109
275
413
575
641
738
882
978
989
1143
1188
1216
1228
47
159
162
159
168
230
242
406
424
Percent Mitt
Eliminator
Restriction
Z
5-7
7-9
7
7
3
0
0
2
1
<1
<1
<1
<1
-
-
<1
-
40
-
-
3
20
-
-
10
-
2
K)
O
o
(a) Intermittent, sequential top wain with makeup water at 0. 55-0. 83 gpm/sq ft for 3 min/hr/section.
(b) Continuous, full face bottom wash with duluted clarified liquor at 0.4 gpm/sq ft.
(c) Intermittent, full face bottom wash with makeup water at 1. 5 gpm/sq ft for 4 min/hr.
(d) A HmeHtone depletion run ia conducted without limestone addition during SOp absorption. The
scrubber inlet liquor pH ia allowed to drop from about S. 9 to 4, 8,
(e) SO_ removals are for 2500 to 3500 ppm inlet gaa SO^ concentration.
-------
Subsequent venturi/spray tower Runs 704-1A through 708-1A confirmed
the observation that reliable mist elimination operation could only be
obtained at high alkali utilization (greater than about 90 percent), with
intermittent underside and topside wash using raw water. Runs 709-1A
through 711-IB also showed that for utilization less than about 80
percent, continuous bottomside wash with diluted clarified liquor could
reduce or limit the amount of soft solids deposition on the mist eliminator
vanes.
Testing with the TCA system similarly confirmed the strong effect
of limestone utilization on mist eliminator reliability. Runs 562-2A,
562-2B and 563-2A (see Table 6-2) were conducted at average scrubber
inlet pH's of 5. 7 to 5. 9 with limestone utilizations from 59 to 71
percent. The mist eliminator was washed intermittently with fresh
water on the topside, and continuously with diluted clarified liquor
on the bottomside. The mist eliminator restriction increased to 7-9
percent during the first 500 hours of operation and appeared to level
out at 7 percent restriction after 811 hours at the end of Run 563-2A.
Following these tests, the scrubber inlet pH was dropped to 5. 2
(Run 564-2A) and the operation continued for additional 113 hours. At
the end of Run 564-2A the mist eliminator restriction decreased to
3 percent from the 7 percent at the start of the run. The limestone
\rtilization during Run 564-2A averaged 94 percent.
As was discussed in Section 5, operation at reduced scrubber liquor
pH and high utilization inherently causes a reduction in SC>2 removal
efficiency. This reduction in efficiency can be compensated by:
(1) increasing the slurry flow to the absorber, (2) increasing the TCA
201
-------
packing height and gas phase pressure drop, or (3) adding MgO to the
scrubber slurry. Another way of increasing SC>2 removal efficiency
while maintaining high limestone utilization is to operate with three
hold tanks in series. TVA is presently conducting an economic evalua-
tion of the proposed schemes for improving SC>2 removal efficiency
while maintaining high limestone utilization.
TCA Runs 565-2A through 580-2A were operated with three hold tanks
in series (see Table 6-2). As with the venturi/spray tower system,
continuous bottomside wash with diluted clarified liquor and intermittent
topside wash with raw water limited the amount of soft solids deposition
at low utilization (Run 567-2A). Also, the mist eliminator was kept
completely free of solids with continuous bottomside wash at high
utilization (Runs 565-2A and 566-2A). As expected, with intermit-
tent bottomside and topside wash at high alkali utilization (Runs 568-2A
through 573-2A),the chevron mist eliminator remained essentially free
of solids. After a total of 1188 operating hours with continuous and
intermittent bottomside wash and intermittent topside wash, the chevron
mist eliminator in the TCA was less than one percent restricted with
soft solids (see Run 573-2A in Table 6-2).
A further example of a mist eliminator which has become less restricted
during operation at high utilization can be seen in Runs 577-2A through
582-2A. At the conclusion of Run 577-2A the mist eliminator was 20
percent restricted by soft solids and after an additional 226 hours of
operation at 95 and 87 percent utilization (Runs 579-2A and 581-2A),
the mist eliminator restriction decreased to 2 percent.
202
-------
6.3 CONCLUSIONS
The following general conclusions can be drawn for the performances
of the single stage, 316 stainless steel, 3-pass, open-vane, chevron
mist eliminator in both the TCA and venturi/spray tower systems
with lime and limestone:
• The reliability of the mist elimination system is a strong
function of alkali utilization.
• For high alkali utilization (greater than about 85 percent),
the mist eliminator can be kept free of solids deposits with
intermittent fresh water top wash combined with either
intermittent fresh water bottom wash or continuous bottom
wash with diluted clarified liquor. Intermittent top and
bottomside fresh water wash may be required in closed liquor
loop operation due to restrictions in the allowable raw water
makeup to the scrubber system.
• For alkali utilization less than about 85 percent, intermittent
top and bottomside wash with fresh water does not limit solids
accumulation. However, for these conditions, a continuous
bottom wash with diluted clarified liquor used in combina-
tion with an intermittent topside fresh water wash can limit
soft solids buildup to less than 10 percent restriction within
the mist eliminator.
• Operation for a period of time at high alkali utilization (greater
than about 90 percent) may, in certain instances, clean up
an already fouled mist eliminator.
The effect of high alkali utilization on decreasing soft solids retention
within a mist eliminator should have direct application to commercial
scrubber design and operation. It may be possible to change operating
conditions in existing scrubbers to increase limestone utilization. High
utilization can be designed into new scrubber systems.
203
-------
Section 7
REFERENCES
1. Bechtel Corporation, EPA Alkali Scrubbing Test Facility; Summary
of Testing through October 1974, EPA Report 650/2-75-047,
June 1975.
2. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program-First Progress Report, EPA Report 600/2-75-050,
September 1975.
3. R. H. Borgwardt, "Increasing Limestone Utilization in FGD
Scrubbers, " presented at the Sixty-Eighth Annual Meeting of
the AIChE, Los Angeles, November 16-20, 1975.
4. R. H. Borgwardt, "IERL-RTP Scrubber Studies Related to Forced
Oxidation of Sludge, " presented at the IERL Symposium on Flue Gas
Desulfurization, New Orleans, March 8-11, 1976.
204
-------
DUQUESNE LIGHT COMPANY
ELRAMA AND PHILLIPS POWER STATIONS
LIME SCRUBBING FACILITIES
R. Gordon Knight, Superintendent, Technical Services
Steve L. Pernick, Jr., Manager, Environmental Affairs
Duquesne Light Company
Pittsburgh, Pennsylvania
ABSTRACT
The Elrama and Phillips Power Stations are coal burning facilities
having net generating capabilities of 494 Mw and 387 Mw, respectively.
Both plants are located in the southwest portion of Pennsylvania within
a radius of approximately 15 miles of the city of Pittsburgh.
Venturi scrubbing systems using lime were installed at both
facilities for dust removal, with one of the scrubbers at the Phillips
facility equipped with a dual stage venturi scrubber to be used and
tested as the prototype for sulfur dioxide removal. Startup of a
portion of the Phillips scrubber system began July, 1973. This system
is designed to scrub the headered gases from all six boilers of the
plant, and discharge them to a new 340 ft. stack after reheating them
from 120° to approximately 140 with direct fired No. 2 oil. The
scrubber system consists of four trains. The first train is the prototype
for sulfur dioxide removal and consists of two identical venturi
scrubbers in series. Each of the remaining three trains consist of
a single venturi scrubber. A series of tests were conducted at this
facility to determine the most feasible means of complying with the
SO emission limitations.
Early operation of the Phillips scrubber system revealed several
problems, such as equipment erosion and corrosion, acid condensate
leakage in the new stack, and stress problems in the ID fans, which
caused outages of the scrubber system. After an extended outage,
the scrubber system resumed partial operation in March, 1974, and has
been in continuous operation since that time with a varying number
of boilers connected to the scrubber system and with a varying number
of scrubber trains in service. For approximately a five month period
in 1975, all six boilers were connected to the scrubber system. During
this period, the scrubber availability was approximately 67%.
Our operating experiences have revealed many problems, such as
scaling and deposits in the scrubber, ID fan deposits and high stresses,
205
-------
a crack in an ID fan shroud, recycle pump erosion/corrosion, high water
inventory, sludge disposal problems, and lime addition problems.
In August, 1975, the scrubber problems were adversely affecting
the plant generating capability to the point where it became necessary
to remove the largest boiler, No. 6 Boiler, from the scrubber system
and return it to its original gas path. Operation has continued with
this arrangement until the present time.
Because of the serious nature of the problems with the Phillips
scrubber system, startup of the Elrama scrubber system was delayed
to permit study and resolution of the problems at Phillips. The Elrama
scrubber system, consisting of five single stage scrubber trains serving
four boilers, was placed in partial operation with one boiler connected
to it on October 26, 1975» A second boiler was connected on February 4,
1976, and operation has continued in that manner to the present time.
Tests conducted on the Phillips scrubber system indicate SO^
removal efficiencies using high calcium lime of approximately 90-6 with
dual stage scrubbing and 50% with single stage scrubbing. Tests using
a magnesium modified lime (magnesium oxide of 8-10%) indicate removal
efficiencies of approximately 83% with single stage scrubbing. The
results of the magnesium modified lime tests indicate an ability to
comply with the S02 emission limitations with single stage scrubbing.
More than $66 million has been spent to date on both scrubber
systems with an additional $20 million to be spent within the next
several years to complete the installations for removing sufficient
SO to comply with the regulations. Operating costs of the Phillips
Station have increased approximately 35% due to the operation of these
scrubbers.
Although many problems have been and are being resolved, others
still remain, and consequently, an assessment of the entire system
operation, reliability, and practicability is inconclusive at this time.
206
-------
DUQUESNE LIGHT COMPANY
ELRAMA AND PHILLIPS POWER STATIONS
LIME SCRUBBING FACILITIES
INTRODUCTION
Duquesne Light Company, an investor-owned electric
utility, serves about one-half million customers in Southwestern
Pennsylvania and has a net generating capability of approxi-
mately 2500 Mw. This capability is generated by combustion tur-
bines in simple and combined cycle modes, by nuclear plants, and
by coal fired power stations. The Company is sole owner and
operator of three coal fired stations, two of which have been
retrofitted with wet scrubbers using lime as a reagent.
The history, description, and early operating experi-
ences of the scrubbers at our Phillips Power Station were de-
scribed in a paper at the EPA Flue Gas Desulfurization Symposium
in November, 1974. A second paper presented to APCA in June,
1975 described the Phillips operating experiences during the
period March, 1974 through April, 1975. This paper will update
the Phillips experiences, giving results of tests that have been
and are being conducted, and it will describe the startup and
operating experiences of the Elrama Power Station scrubbing sys-
tem for the period extending through January, 1976.
DESCRIPTION OF SYSTEMS
Phillips Power Station
Four trains of wet venturi-type scrubbers have been
installed at the 387 Mw Phillips Power Station at a cost of ap-
proximately $103 per Kw. Gibbs & Hill, inc. was engaged as
architect-engineer for the entire project, with battery limits
of the Chemico Corporation confined to the scrubbers and asso-
ciated pumps and controls between the inlet hot gas duct mani-
fold and the exit wet gas header, including the reheater but
excluding the new induced draft fans.
The four trains are located downstream of existing in
series mechanical collectors and electrical precipitators in-
stalled on each of the six pulverized coal fired boilers. Three
of the trains are single stage venturi scrubbers originally in-
tended for particulate removal. The fourth train is dual stage
and was the prototype for determining the feasibility of two
stage scrubbing for compliance with S02 emission limits. A test
program on the prototype and the single stage scrubbers was used
to determine what additional equipment or process will be neces-
sary to enable us to meet regulatory requirements for S02 re-
noval, which are essentially 83% S02 removal efficiency using 2%
sulfur coal.
-------
Each train is equipped with a new wet-type ID fan. A
new common duct directs emissions from all boilers to the scrub-
ber plant where they can be sent to any or all of the trains.
The scrubbed gas is exhausted through a common duct and an oil
fired reheater to a new ground supported stack which consists
of a concrete envelope, a 30 inch annulus, and an inner acid-
resistant brick stack.
Slaked quicklime is added to the lower cone of each of
the scrubber vessels to neutralize the recirculating liquor,
which, in single stage scrubbing with high calcium lime, reacts
with about 50% of the S02 in the flue gas. A liquor bleed flow
of about 4% is sent to one or both of two thickeners for solids
removal. The overflow is returned to the system, and the under-
flow is piped to the sludge treatment system.
Startup of a portion of the Phillips scrubber system
began July, 1973. Several problems then developed causing out-
ages of the scrubber system, and after an extended outage, the
scrubber system was returned to service in March, 1974. The
system has been operating continuously since that time with a
varying number of boilers connected to the scrubber system, and
with a varying number of scrubber trains in service.
Elrama Power Station
The scrubber facility at the 494 Mw Elrama Station is
identical to that at Phillips in most ways. Gibbs & Hill is the
architect-engineer, Chemico scrubbers have been installed, and
the battery limits are the same. Mechanical and electrical dust
removal equipment remove most particulates from the boiler emis-
sions, and the gas to and from the scrubbers is headered in the
same way. Unlike Phillips, there are four boilers, each with
its own turbine generator. Five single stage scrubbers were in-
stalled, and it was intended that the knowledge gained from the
test program at Phillips would be applied to Elrama to enable
compliance with the regulations.
The first Elrama scrubber was not placed in service
until October 26, 1975. It had been scheduled for an earlier
startup date; however, the severity and number of problems en-
countered at Phillips made it prudent to delay startup until
many of the problems at Phillips were resolved and the necessary
modifications made at both stations. The decision to delay the
Elrama startup has proved to be a wise one.
PHILLIPS OPERATING EXPERIENCE
A number of problem areas were described in our previ-
ous papers. Some of these have been alleviated, and others
still persist. The present status of a number of these is de-
scribed below.
208
-------
ID Fans
Encouraging progress has been made in this area. The
second repair of the original defective welds with Inconel 112
has proved successful and resulted in very few minor or no weld
repairs at each 1500 hour inspection. Relocation of spray noz-
zles with the use of a new type (Bete fog nozzle No. TF16FC) has
practically eliminated the deposit accumulations, which caused
fan imbalance and under which pitting attack occurred.
A problem still exists with pitting of the 316 SS
sleeves on the fan shafts. Recent trial use of a coating of
Neoprene-29 appears encouraging. The rubber lining on the hous-
ings needs periodic minor repairs.
During a routine inspection of a fan in the last week
of January, a crack was found in one of the shrouds. The crack
extended about 4 inches from the outer periphery and was about
1/16 of an inch wide at that point. Both shrouds will be
checked by non-destructive methods. The defective area will be
removed for examination. Repairs will be made, and the fan re-
turned to service. This fan has experienced about 11,000 serv-
ice hours. The other fans will be inspected and checked by non-
destructive methods.
Recycle Pumps
The corrosion/erosion problem with these pumps con-
tinues. As indicated in our previous paper, we are following a
trial program involving a number of different impeller and wear
ring materials. Although we have still not obtained the desired
life from any of the materials we have tried, there are indica-
tions that a CD4MCu impeller with a 26% chrome-iron wear ring is
the best combination. Service life with these two materials has
been up to 4700 hours, compared to the 3000 hours with the or-
iginal Carpenter 20. Two rubber lined pumps are on hand for
trial installation at the Elrama Station.
Bleed Valves
Previous problems with rubber lined plug-type bleed
valves have apparently been solved by the substitution of pinch-
type valves.
Deposits
Experience has shown that complete cleaning of a
scrubber vessel is required about every 1400 service hours.
Each cleaning with minor maintenance requires 1400-1700 man
hours. Approximately 20-80 tons of deposits are removed during
each cleaning. These deposits result even though the pH of the
recycle liquor is maintained in the 6-7 range for about 95% of
the time. We do not yet have redundant lime feed or automatic
pH control. These are to be forthcoming.
209
-------
The main troublespot for deposits is in and around the
throat dampers. Despite programmed, automatic exercising of the
dampers/ deposits accumulate to the extent that the throat pres-
sure drop increases from a normal 6 inches to 12 inches or more
after about 1400 hours of service. This high delta p decreases
the amount of gas that can be scrubbed to such an extent that
generating capability of the power plant is reduced and a clean-
ing outage is necessary. The scrubber manufacturer had in-
stalled and recommended that we keep open a 1-1/2 inch valved
vent on each side of the damper housings in an effort to sweep
the areas clear of deposit with outside air. This had no ef-
fect. To obtain more air sweep, we next removed eleven 3/4 inch
plugs from openings on the sides and tops of four of the twelve
damper housings on one of the vessels. This, too, had no ef-
fect. Earlier we had introduced a continuous flow of water
through the same 3/4 inch openings with little or no benefit.
Deposits accumulate in and around the twelve tangen-
tial nozzles in each vessel. This results in an uneven flow of
scrubbing liquor around the upper cone. Chemico recommended
that alternate nozzles in one of the vessels be blanked off to
see if the increased velocity through the others would prevent
the accumulations. After 1506 service hours, this appears to
be partially effective in the vessel where it was tried.
Sizable deposits form in the lower cone, where the
lime discharges into the vessel through an open end pipe.
Chemico recommended the use of an angled nozzle at this loca-
tion. This is currently being implemented.
Appreciable deposits are also found on the lip under
the upper cone. Chemico is conducting scale model studies in
an effort to determine the cause and possible corrective meas-
ures.
Comments on all these deposit problems should be con-
sidered in conjunction with the results obtained in our magnes-
ium modified lime tests. These will be discussed later.
PHILLIPS SLUDGE TREATMENT AND DISPOSAL
Previous reports have discussed our first-of-a-kind
experience with Dravo's Calcilox as a fixating agent. Early un-
successful use resulted from our inability to maintain one or a
combination of (1) the required 10% addition (by dry weight of
sludge), (2) warm ambient temperature, (3) necessary curing
time, and (4) pH of 10-11. None of these would be critical if
we could add the fixative and then let the mixture remain in
place. However, we have only three "curing" ponds, each of
about 6000 yd3 capacity, and we fill each one in 10-14 days.
Furthermore, to be prepared for unexpected emergencies, such as
the draining of a scrubber or a thickener, we must have an empty
or near empty pond at all times. Consequently, we must excavate
210
-------
ponds within about 4 days of filling which precludes the neces-
sary curing time. The problems involved with hauling "Lup" in
open dump trucks are obvious. When requirements are satisfied,
y^r^aiV£;e 11S5* fixation with Calcilox. The original system
introduced the additive in slurry form. In April, 1975, we in-
nravi ?h*ry add*tlve system, fabricated by and rented from
wSh ™« J WaS lnt
-------
PHILLIPS SCRUBBER SYSTEM AVAILABILITY
Reports by others on availability of the Phillips
scrubbers have been hopelessly confused and unduly optimistic
because of the header design of the plant. Until all boilers
were connected to the scrubbers, there was at least one spare
train and as much as 100% spare scrubber capacity. Therefore,
having a train out of service for maintenance did not reduce
the capability of the scrubber plant. Until all six boilers
were connected, meaningful availability factors were not avail-
able. On March 17, 1975, the last boiler was connected, and
all four scrubber trains were required to be in service. Opera-
tion in that mode continued until August 4, 1975 when it was
necessary to remove the No. 6 Boiler from the scrubber system
because the pH in the system could not be maintained, and depos-
its were becoming unmanageable to the point where scrubber out-
ages were reducing the station generating capability. In the
four and a half month period of total scrubbing, availability
averaged 67%.
PHILLIPS SCRUBBER SYSTEM TESTS
In addition to our continuing performance and sludge
tests, several other tests are noteworthy.
Lamella Thickener
Because of the space limitations at both stations and
because of the success of the phosphate industry in using the
Lamella plate-type thickener, we obtained and operated a 100 gpm
model at Phillips for four weeks. A polymer was added to the
inlet stream from the scrubber bleed. The influent contained
2.5% to 3.0% solids, and the Lamella increased the concentration
to the range of 17-20%. We require a 30-40% concentration. The
overflow quality was excellent - less than 150 ppm. The manu-
facturer said that use of a larger bottom hopper would have
given a greater concentration of solids in the underflow. No
good evaluation of long term deposit accumulation and ease of
removal can be given. The Lamella plates were disassembled
after 17 days of continuous operation, and soft deposits of 1/4
to 1 inch in thickness were found near the bottom of the plates. .
These were easily removed from the plates, after removal from
the housing, by medium pressure jets. Removal of deposit from
in-place plates cannot be evaluated.
Additives
We have been plagued with black plumes from the scrub-
ber stack at times of boiler upsets. Although we were originally
told that the scrubbers would remove unburned carbon, this has
not been the case. In an effort to clean up the plume, we tried
four different wetting agents, all from the same supplier. None
had any effect.
212
-------
Magnesium Modified Lime
We ran a two and a half month test with magnesium
modified lime in order to determine (1) if single stage scrub-
bing could remove sufficient SC>2 to enable compliance with
regulations, (2) the effect on deposits, and (3) the change in
sludge characteristics. Because of our limited lime capacity,
it was necessary to reduce the quantity of emissions treated in
the scrubbers. We did this by re-routing the emissions from a
second boiler back to the original stack - this boiler was later
taken out of service for major maintenance - and by limiting op-
eration on a third to supplying emergency generation require-
ments. In order to slake the blended lime completely, a source
of hot water was obtained so that slaking temperature was held
in the 180-190°F range. Efforts were made to reduce the addi-
tion of fresh water in order to maintain system chemistry. An
effective step was the use of thickener overflow return water
to dilute the lime slurry after slaking. A not-so-effective
step was the reduction of demister spray time. Deposits re-
sulted until normal spray time was restored.
System requirements limited the duration of the test
so that the optimum value of some operating parameters must be
determined at a later time. However, the test did yield some
valuable information.
It was determined that the required degree of S02 re-
moval - approximately 83% - can be obtained with an MgO content
of 8-10% in the lime with single stage scrubbing. This will
eliminate the necessity to install any additional stages or
scrubber vessels at both the Phillips and Elrama Stations. This
will save approximately $10 million in capital costs.
The two and a half month test with modified lime re-
sulted in 1612 service hours on one train and 1309 on another.
After this length of service, the throat damper pressure drop
in both trains was still 6 inches, the same as with newly
cleaned dampers. With the normal high calcium lime, the pres-
sure drop would have been at or approaching 12 inches, and an
outage for cleaning would have been necessary. Inspection
showed negligible deposit accumulation in all accessible loca-
tions, and the deposit on some previously scaled surfaces had
been removed. For example, the maximum flow that could be ob-
tained through the bleed line on one scrubber prior to the test
was 420 gpm. This flow increased during the test to a value of
720 gpm at termination. On the basis of SO2 removal efficiency
and freedom from deposits, it is unfortunate that we cannot con-
tinue the use of the modified lime. As discussed later, this
requires major modifications, about 18-24 months for construc-
tion and the expenditure of about $20 million to convert both
the Phillips and Elrama scrubbers.
One of the less favorable effects of using magnesium
Modified lime was on the sludge, which did not settle as well in
213
-------
the thickeners. The solids concentration in the underflow de-
creased from 35-45% to 25-30%. Although a polymer was tried,
time did not permit a complete evaluation of its effectiveness.
The nature of the solids is important from two considerations:
effect on sludge treatment schemes and on the size or number of
additional thickeners needed for the next phase of construction.
Steps were taken to supply sludge samples to the thickener
builders and to the three bidders for our long term waste man-
agement contract. The results of the thickener tests have not
yet been submitted to us. All three of the waste processors
said that the sludge could be handled, with slight modification
needed in some cases.
Another unexpected effect of our use of magnesium
modified lime was a reduction in the pH of the sludge (8-9 in-
stead of 10-11}. This evidently resulted from better lime slak-
ing, resulting in less unreacted lime in the sludge, and from
higher utilization of lime by better S(>2 removal. The effect on
sludge fixation was to leave us with soupy ponds to excavate and
haul to the final disposal site. Dravo recommended the use of
15% (instead of 10%) Calcilox addition, and they added an "alka-
line booster" to the Calcilox. We were not able to get the de-
sired pH before the test ended, and, to make things worse, the
increased SC>2 removal and consequent sludge filled the ponds
faster and reduced the curing time. Dravo's representatives in-
dicated that laboratory tests showed good fixation with the
proper pH and curing time.
ELRAMA SCRUBBER STARTUP AND OPERATING EXPERIENCE
As previously noted, the first boiler was connected to
the scrubbers on October 26, 1975. This boiler, No. 2, has an
equivalent capacity of about 100 Mw, and the emissions can be
handled by one scrubber. However, to insure reliability in case
of a scrubber .malfunction, two scrubbers are operated at partial
load to protect the boiler and its associated turbine generator
against a trip-off. We have had continuous operation of the
boiler on the scrubber system through January, 1976 with the
exception of two short outages. One was occasioned by inopera-
tive throat dampers, originally thought to be a result of depos-
it buildup. Upon inspection, however, the cause was found to be
damper scraper blades which had been installed with insufficient
clearance. The other involved failure of the lime feeder belt
in conjunction with a boiler outage. Four of the scrubber ves-
sels have been in service in various combinations, and the serv-
ice hours to January 31, 1976 are:
No. 1 - 1169 hours
No. 2 - 1508 hours
No. 3 - 976 hours
No. 4 - 838 hours
No. 5 has not been in service because it is being revised for
the trial installation of rubber lined recycle pumps.
214
-------
To this time, there have been no major problems.
There have been many startup pains, not the least of which has
been frozen pipes. Heat tracing has been and continues to b,i
a problem at the Phillips Station, and it appears we have more
of the same at Elrama. The most persistent problem, and one
that could take the scrubber system and generating facilities
out of service, has been experienced with the thickeners. This
is a hardware and design problem associated with recirculation
of the sludge within the thickeners to attain the 30-40% solids
concentration. Fortunately, station personnel have recently
found that our sludge treatment process (described below) can
handle sludge with solids concentration as low as 15%; thus.
recirculation is unnecessary. Each of the two thickeners has
sufficient capacity for one boiler. Therefore, we have had
spare capacity up to this point. When another boiler is con-
nected to the scrubbers, a thickener outage will cause a boiler-
turbine generator outage. Two boilers will be the total that
can be placed on the scrubber system until additional construc-
tion is completed.
ELRAMA LIME SYSTEM
A report appearing in the literature in August, 1975
indicated that the Phillips scrubbers were the first in the
United States to go in service using hydrated lime. That was
a misstatement. However, it can now be said that the Elrama
scrubbers were placed in service and are continuing with hy-
drate. However, we will change over to quicklime as soon as
possible since it is more economical than hydrated lime.
The installed lime slaker, silo, and pumps had a feed
capacity of 9 Ib/minute with all scrubbers in service. Since we
learned through experience at Phillips that this was inadequate
for proper pH control, we did not want to go through a similar
problem at Elrama, and because procurement of additional equip-
ment would have caused an impossible delay, steps were taken to
adapt the system to hydrated lime. The slaker was converted to
a mixer, the silo feeder changed, and new transfer pumps in-
stalled. In addition, arrangements were made for the lime sup-
plier to leave two loaded 20 ton trailers on site at all times.
These, along with our 20 ton silo and a recently purchased 50
ton storage "blimp" and permanently installed lime blower, gave
us a total storage capacity of 110 tons and a feed rate of about
90 Ibs/minute. The system has worked well, and a pH of 7-7.5
has been maintained with about 30 tons/day consumption for the
one boiler. This is the equivalent of about 30 Ibs/minute of
quicklime.
ELRAMA SLUDGE TREATMENT AND DISPOSAL
Under the original concept, it was believed that only
fly ash would be removed in the single stage scrubbers. The un-
fixed sludge would be dewatered in three clay lined ponds, which
215
-------
would be excavated as required, and the sludge placed on our
normal landfill. As we and the industry now know, there is al-
ways some SC>2 removal, and with the pH we consider necessary,
the quantity of SC>2 removed is not negligible. And, as we have
sadly learned, unfixed sludge is not easily excavated or trans-
ported. For this particular situation and with our recently
completed test of the IUCS prototype at Phillips, the POZ-O-TEC
system seemed to be the answer to our dilemma. Fortunately,
IUCS had just completed their tests at the Mohave Station, and
that equipment was available. After entering into a contract
with them, the equipment was moved from Nevada, and they were
ready to process sludge before we were ready to produce it.
Shortly after startup, they were able to obtain and install a
vacuum filter which increases the 35-40% solids in our underflow
to 50-60%. Consequently, their process requires less dry fly
ash to be mixed with the sludge.
COSTS
Thirty-two million dollars has already been spent on
the Phillips scrubber system with the expectation of an addi-
tional $8 million to be spent for additional equipment to come
into compliance with the SO? emission limitations for the entire
plant. Approximately $34 million has already been spent on the
Elrama scrubber system with the expectation of an additional $12
million to be spent for final compliance with the SC>2 emission
limitations. These expenses equate to an installation cost of
$103 per Kw for the Phillips Power Station and $93 per Kw for
the Elrama Station.
The actual operating costs we have been incurring for
the Phillips scrubber system, which is not yet complete, are ap-
proximately $9 million per year and are equivalent to approxi-
mately 5 mills per Kwhr, 38* per million BTU, or $8.50 per ton
of coal consumed, all including fixed charges. Based on these
costs, the combined operating costs for Elrama and Phillips when
full SO2 compliance is achieved will be approximately $32 mil-
lion per year, or 6 mills per Kwhr, or 52$ per million BTU, or
$12 per ton of coal consumed, all including fixed charges.
The estimated disposal costs of sludge are approxi-
mately $6 to $7 per wet ton, or $12 to $14 per dry ton. These
costs may change upon selection of a contractor for long term
disposal and upon determination of the cost of acquiring and
developing additional disposal areas.
We expect that the costs of operating these scrubber
systems, when completed, will result in an increase in the cus-
tomers' bills by approximately 10% over 1974 billing levels.
FUTURE PLANS
Our next phase of construction toward total scrubbing
for SO2 removal at Phillips and Elrama has been known in our
216
-------
Company as Phase IA. However, with the now recognized impor-
tance of continuous pH control, we have gone into an accelerated
Phase IA. This is intended to give us redundant lime feed at
the Phillips Station by May, 1976, for more reliable scrubber
operation, and adequate quicklime feed at the Elrama Station for
scrubbing the gas from two boilers by June, 1976.
Completion of the total Phase IA program will provide
the lime feed and storage facilities, thickeners, and auxiliary
equipment sufficient to enable compliance with SC>2 emission reg-
ulations. This will include automatic pH control with the re-
dundancy required for continuous control. This phase should be
completed by December, .1977 at both stations. We plan to use
magnesium modified quicklime, not only to achieve compliance
without the necessity of installing additional vessels, but also
to avoid the deposits, which we experience with high calcium
lime.
Sludge treatment and disposal is still an unsolved
problem in our opinion. Despite glowing reports of presently
available methods, we are not convinced. We probably have more
experience with the available methods than any other scrubber
operator, and none of them has yet been demonstrated as being
environmentally acceptable. As mentioned previously, we are in
the process of receiving proposals for long term sludge disposal
from three different organizations. We have specified that the
process chosen must yield a "structurally stable and environ-
mentally acceptable" product. We want all this and a process
that we and our customers can afford. As with all scrubber de-
velopments, this should prove interesting - and, no doubt, frus-
trating.
CONCLUSION
In previous papers we had listed what we feel are
levels of performance that must be satisfactorily resolved if
the system is to be considered operationally feasible. With
regard to our experience to date, we presently assess these per-
formance levels as follows:
1. Reliability - Although we have experienced un-
acceptable reliability with the use of normal
high calcium lime at our Phillips Station, our
test program using magnesium modified lime in-
dicates that an improved reliability can be
achieved. In addition, resolution of several
of the problems at Phillips and the implemen-
tation of those solutions at the Elrama fa-
cility prior to initial operation indicate
that some improvement may be achieved in the
reliability of the Elrama scrubbers. However,
the Elrama system has not yet been run under
full load conditions which would more accur-
ately reflect true reliability.
217
-------
2. Turndown Capability"- Over the course of opera-
tion in the past year, it has been possible to
follow the plant load with little problem pro-
vided sufficient scrubber capacity is available.
However/ we have not been able to determine if
such cycling operation contributes to the scal-
ing and deposit problems that have been experi-
enced .
3. Closed Loop Operation - Our goal is to achieve
closed loop operation, however, this appears to
be several years away. Some improvement was
made in reducing blowdown, but it appears that
further improvement will have to wait for the
use of magnesium oxide lime and the addition of
more lime feeders and associated equipment.
4. Sludge Disposal - Our experiences with several
sludge disposal techniques indicate that such
techniques can result in acceptable stabiliza-
tion of the sludge. However, there is insuf-
ficient information on long term leaching
characteristics of this material to claim such
techniques as being environmentally acceptable.
5. Cost Consideration - Our installation and operat-
ing experiences with the Phillips scrubber system
indicate that upon completion of the Elrama and
Phillips scrubbers, the operating costs for these
two scrubbing systems would increase the customers'
bills by approximately 10% over 1974. Although
this is not a physical performance level, the
economic aspect must be considered in a total
system evaluation.
218
-------
OPERATIONAL STATUS AND PERFORMANCE OF THE COMMONWEALTH
EDISON COMPANY WILL COUNTY LIMESTONE SCRUBBER
Warren G. Stober
Commonwealth Edison Company
Chicago, Illinois
ABSTRACT
The Will County Scrubber was installed in 1972 as a full scale
demonstration plant. Over the past four years numerous problems have
been encountered, many of which have been solved. However, reliable
operation on high sulfur coal has not been achieved. This paper
reviews the past operating history, some of the major problems and what
has been done to solve them, scrubber availability, sludge treatment
operation, the cost to own and operate the scrubber system, and
future plans for Will County.
219
-------
OPERATIONAL STATUS AND PERFORMANCE
OF THE COMMONWEALTH EDISON COMPANY
WILL COUNTY LIMESTONE SCRUBBER
In the spring of 1970, Commonwealth Edison contracted
with Bechtel Corporation to investigate the sulfur removal
systems then available and to help us select a system having
a sufficient degree of development to warrant a large scale
installation on our Will County Station's Unit #1.
After deciding upon a wet scrubber system using lime
or limestone, a specification was then prepared by Bechtel
and released for bid. From the nine bidders that were
solicited, seven proposals were received. After detailed
study and bid evaluation with special consideration of the
project schedule, Babcock and Wilcox was given authorization,
in September, 1970, to begin the detailed engineering for a
limestone slurry system. The project had a completion
deadline of December 31, 1971 which was established by the
Illinois Commerce Commission as part of a rate case.
Design
The Babcock and Wilcox designed process was guaranteed
to remove 98% of the f" / ash and 76% of the sulfur dioxide.
These efficiencies were Msed on a dust inlet loading of
1.355 grains per standard ".ubic foot at 70°F and burning 4%
sulfur Illinois coal.
The Will County wet ^rubber was backfitted on a
177,000 gross kilowatt Babcock and Wilcox radiant cyclone
boiler that was put in service in 1955. The location of the
scrubber and its associated equipment is indicated on Figure 1,
Property Plat of Will County Station.
Construction presented a great many problems both
physically and schedule-wise. It was necessary to sandwich
the scrubber between the boiler house and the service building
with a substantial cantilever. Erection of equipment began in
May, 1971 and with the use of substantial overtime, the
majority of the system was completed by February, 1972. The
tight retrofit and quick schedule resulted in a very expensive
system: $108 per net kilowatt in 1972 dollars including
indirect costs.
220
-------
The wet scrubber system is divided into three parts:
the limestone milling system, the wet scrubber and absorber,
and the sludge disposal system.
The milling system as shown on Figure 2 consists of a
limestone conveyor, two 260-ton capacity limestone storage
silos, two full-sized wet ball mills, each designed to
pulverize 12 tons of limestone per hour so that 95% will pass
through a 325 mesh screen, and a slurry storage tank. The
required limestone is high in calcium carbonate, above 97%,
and low in magnesium carbonate, less than 1%.
The wet scrubber is made up of two identical systems,
each taking half the boiler flue gas. Each module is designed
for 385,000 cfm at a gas temperature of 355°F. Each system
consists of two recirculation tanks, venturi and absorber
slurry recirculation pumps, a venturi fly ash scrubber, a
sump, a two-stage sieve tray sulfur dioxide absorber, a two-
stage mist eliminator, flue gas reheater, and an ID booster fan.
Each system contains a by-pass damper which permits a module
to be taken out of the gas path. Figure 3 illustrates the
entire system showing the relative location of each of the
above components.
Figure 4 illustrates the flue gas path. The flue gas
emerges from the boiler, passes through an existing
electrostatic precipitator and on into the venturi. The flue
gas velocity at the throat is approximately 130 feet per second
with the pressure drop being maintained at approximately nine
inches of water. From the venturi the gas flows goes through
the sump and then upwards into the absorber. Here the sulfur
dioxide is removed as the gas is forced through two separate
sieve trays at a greatly reduced velocity, about 10 feet per
second. From the absorber the gas passes through a reheater
and on to the booster fan.
Figure 5 illustrates the slurry recirculation system.
The flow of slurry to each venturi is 5800 gallons per minute.
and to each absorber 11,200 gallons per minute. This gives a
liquid to gas ratio of about 18 to 1 and 35 to 1 respectively
at full load.
The waste slurry is pumped to either a thickener or to
back-up settling ponds and all the clarified water runoff is
recycled to the scrubber and milling system.
221
-------
The power requirement to operate the scrubber system
is approximately 7,000 kilowatts or about 4$ of the unit
gross capacity of 177,000 kilowatts. Over and above this
power requirement is a need for steam to reheat the gas
exiting the system. The system was designed to reheat the
gas from 128°F up to 200°F, using about 50,000 pounds per
hour of steam at 435°F and 350 PSIG.
Performance 1972-1974
B-module was the first module to come on line on
February 23, 1972, and wsis followed by A-module coming on line
on April 7, 1972. Both modules were plagued by more problems
than one would expect to be normal during start-up. Demister
fouling was very prevalent'and numerous changes were made to
the demister wash system in an attempt to solve this crippling
problem. Other major problems that affected availability
were reheater tube vibrations, reheater fouling, scaling on
the absorber walls and grid plates, and slurry nozzle pluggage.
Through the end of 1972, A-module operated 1444 hours,
the B-module operated 1237 hours, thus attaining availabilities
of 29.596 and 25.296 respectively. (Availability, in this case,
is calculated by dividing the operating hours of the scrubber
by the operating hours of the boiler-turbine unit).
Simultaneous operation was 469 hours, for an availability of
only 9.6%. The longest period of continuous operation was 21
days, this being with A-module, and the longest simultaneous
operation of both modules was 6 days.
By early 1973 many problems had been resolved; however,
the major ones remained: demister pluggage, reheater pluggage,
and scaling. Along with the recurring major problems, new
problems were encountered, such as cracks in the booster fan
inlet cones, chloride stress corrosion cracks in the 304
stainless steel reheater tubes and eventual failure of the Corten
reheater tubes.
In April 1973 a decision was made to discontinue
operation of B-module and concentrate all efforts on achieving
satisfactory operation on A-module. If A-module could
demonstrate satisfactory operation, then B-module, which
originally had a turbulent contact absorber, would be modified
so as to be identical to A-module.
222
-------
A-module, after operating for 1726 hours through
November, 1973 (22.6% availability) was taken out of service
for extensive changes in the demister, demister wash system,
and reheater. The underspray water supply was changed from
pond reclaim water to house service water. This change was
made to try to prevent the scaling that was taking place on
the demister. The new second stage demister was designed
to knock out any wash water carryover in the gas flow from
the first stage demister top wash system. The intent of
these changes was to eliminate demister scaling and
carryover into the reheater coils. In an attempt to slow
down the rate of deterioration of the reheater coils, a
steam supply line was run from the building heating system
to the coils for use during scrubber outages. Keeping the
coils hot and dry during outages reduced the corrosion due
to chlorides deposited on the coils during operation.
A-module returned to service in late March, 1974
following completion of the modifications and an additional
delay to repair damage caused by a freeze-up of instruments
and the pond recirculation system.
During 1974 we changed the method of calculating
scrubber availability from the ratio of scrubber operating
hours to boiler operating hours to the ratio of scrubber
available hours to the total hours within a given time
period. While neither method is completely satisfactory,
we believe the new approach provides a more accurate
indication of scrubber performance. In some cases this has
increased the availability because the repairs to the
scrubber were completed before the boiler repairs. However,
for the sake of those who still like to look at scrubber
availability based on boiler operating hours, these numbers
have also been calculated.
In 1974, A-module was available for service 6025
hours for an availability of 68.894. A-module operated 5468
hours and unit 1 was operational 7,924 hours giving a
scrubber availability based on operating hours of 69.0%.
After achieving what seemed like a workable system with
A-module, we decided to proceed with the modification of
B-module, incorporating all the changes made to A-module. The
major changes included the removal of the TCA ping-pong balls
and replacement with two stages of sieve trays, the single
stage demister being replaced by two stages of a sturdier
molded design, and the reheater bundles being replaced with
223
-------
5/8" OD X 0.065" 316L stainless steel tubes in the lower
three banks and similar sized carbon steel tubes in the
four upper banks. B-module remained out of service through
the end of 1974 and well into 1975.
1975 Operation
¥e entered 1975 operating in much the same manner
as we operated in 1974. A-module was operational and B-module
was still undergoing modifications. The first three months
of 1975 saw A-module achieve availabilities of 99-5%, 99.4%,
and 94.0% (based on hours that the module was available for
service). Table I summarizes 1975 availabilities computed
by both methods, one based on boiler-turbine operating hours
and the other based on the hours the scrubber was available
each month.
Figures 7 through 16 are monthly charts indicating
all the unit and scrubber outages. While these sheets give
the general nature of the outage, they do not specifically
mention or detail the work done on the scrubber during unit
outages. It should be noted that it was always necessary to
do some scrubber maintenance work during those outages.
During the first three months of 1975 the^e were
eight A-module outages. However, only one was a forced outage,
the others being for no unit demand, inspections and one
accidental trip. The forced outage was due to the splitting
of a slurry supply hose to the venturi.
At the time the reasons for the increased availability
during 1974 and the early months of 1975 were hard to ascertain.
Looking back today at what took place then and what was to
follow later in 1975, we believe that the type of coal we
were burning at the time had a lot to do with the increased
availability. The addition of a second stage demister,
automation of pH control and other minor changes had some
influence, but were not determining factors. As part of our
commitment to reduce S02 emissions, only low sulfur (approximately
0.4?' S) coal was being shipped to Will County Station. As
this type of coal was mixed with and gradually displaced the
high sulfur coal remaining in the storage pile, the average
sulfur content of coal burned in Unit #1 dropped. In 1974, the
average was 1.5#; in early 1975, the average was slightly over
1.056. We believe that the low sulfur content of the coal and the
change from sludge pond reclaim water to house service water for
the demister wash were the primary causes of the improved
reliability. This conclusion was tentatively confirmed when
we attempted operation on high sulfur coal later in 1975 and
encountered severe problems.
224
-------
The change from sludge pond reclaim water to house
service water for demister wash water upset the water
balance of the system. The high availability of A-module during
the first three months of 1975 which resulted in part from
this wash water change, caused the water level in our sludge
pond to frequently rise to the point where we became concerned
about an accidental overboarding to the nearby Des Plaines
River. In April the scrubber was shut down until changes
could be made that assured overboarding would not occur
Three steps were taken to tighten the water balance and assure
a closed loop operation: 1. Pump gland water flows were
reduced by 50%; 2. The scrubber house service water filter
backwash was routed out of the system; 3. The continuous
demister wash underspray was changed to an intermittent spray
of 5 minutes on - 5 minutes off. A-module returned to service
on May 5, 1975 and the noted changes proved to be effective in
maintaining a lower pond level thereby assuring that overboard ne
would not take place. &
B-module returned to service on May 20, 1975 upon
completion of its modification program, and the start up was
rather uneventful with only minor problems being encountered
which resulted in a couple of short outages.
Early in the year it was decided that we should attempt
to operate the scrubber on Illinois high sulfur coal. Because
Will County was receiving only low sulfur coal as part of our
S02 compliance program, it was necessary for our Fuel
Department to arrange for special shipments of high sulfur
coal for the test. Difficulties in obtaining and storing
the test coal forced us to begin burning it before we had
obtained sufficient experience operating both modules
simultaneously or with the tight water balance which had
Oust been put into effect. As a result, we were not totally
prepared for the myriad of problems which began to occur.
During the high sulfur coal burn program, SO? removal,
particulate removal in the scrubber, and precipitator efficiencies
were measured along with extensive monitoring of cycle chemistry
Removal efficiency tests were conducted under full and partial
load conditions and with the electrostatic precipitator ahead
of the scrubber in and out of service. S0? removal efficiencies
averaged 78.2% on A-module and 86.8% on B-module with inlet
S02 loadings averaging 3573 ppm. A-module removal efficiencies
were lower because only one of the two absorber slurry pumps was
in service, resulting in a lower liquid to gas ratio. The other
absorber pump was out of service for repair of the impeller and
rubber lining.
225
-------
Particulate removal efficiencies were very erratic
and lower than expected. Scrubber outlet loadings, which
were measured using the EPA Method #5 train, were high;
with a large percentage being slurry carryover rather than
fly ash. During the test period slurry solids in the
recirculation tanks rose to a higher level than desired.
In an attempt to reduce the solids content, additional water
was added to the cycle by changing the demister wash cycles.
Bottom wash was made continuous and the top wash frequency
was increased. The top wash is a deluge system, utilizing
a high flow for a short period of time. But since this was
pond return water, the increased use of this system prompted
scale to form on the walls. Scale and pluggage in the demister
apparently resulted in uneven gas distributions and high
velocities which permitted slurry to carryover through the
system.
Ten days after the start of the test an abnormally
high pressure drop was noted across the A-module reheater.
An inspection showed the first stage demister starting to plug
and the reheater coils heavily fouled. B-module was taken out
of service 13 days after the start of the test because of high
vibrations of the booster fan. At that time inspection of the
module showed severe demister pluggage and the fan rotor coated
with a dark scale. Approximately four days were required to
clean and balance the fan before B-module could be returned to
service.
By mid June, 18 days after commencing the high sulfur
coal test program,A-module was out of service due to massive
reheater leaks, reheater fouling and demister pluggage. It
remained out of service for repairs for the remainder of 1975.
With tight control of the water balance in order to
assure a closed water loop, two modules in service and operation
on high sulfur coal, changes within the scrubber cycle were
occurring quicker than we could detect and correct them. With
the turn of events being what they were, it is difficult to
assess the operation or draw any firm conclusions from the
attempted high sulfur coal test other than to ss>y we were unable
to successfully operate on high sulfur coal under these conditions.
Shortly after the test shipment of 20,000 tons of 4%
sulfur Illinois coal was consumed, it was decided that coal from
the storage pile would be burned in Unit #1. Approximately
90,000 tons of coal ranging from 0.8% to k% sulfur remained
226
-------
segregated from the remaining 0.4% sulfur coal in storage
and was used to fuel Unit #1 when B-module was in service.
However, during the months of July, August, and early
September 1975 there were occasions when low sulfur (0.4%)
coal was used to fuel the boiler and this resulted in a large
variation of the inlet S02 concentration in the flue gas
entering the scrubber. By mid September, B-module was taken
out of service because of massive scaling in the absorber,
in the sump area just below the absorber and in the demister.
Presumably this was caused by the pH control either not
functioning properly or the system just not being able to
handle such large variations in the inlet SO- conditions.
An outage of over one week was necessary to nand-chip the scale,
which in some places was one-half inch thick, off the various
internals.
Another major problem that affected performance of
B-module began occurring with less than 1000~hours of operating
time on the rebuilt module. This was the failure of the new
carbon steel reheater tubes. We eventually experienced
failures in six of the twelve new tube bundles over a three
month period. Peripheral cracks occurred in the tubes just
before they entered the header. In all cases where failures
occurred, the tube supports had worked loose permitting the
tube to vibrate. The coil manufacturer has repaired and
modified all the tube bundles by redesigning the tube supports
and their method of attachment to the frame, including an
additional support.
Other parts of the system have also experienced
problems which are worth noting. The horizontal plate above
the nozzles at the venturi inlet wall wash box developed numerous
holes allowing slurry to splash and drip over the scrubber
structure. The cause was found to be slurry being deflected
onto the top plate and eventually eroding through at the point
of impingement. Each nozzle has now been fitted with a deflector
designed by B&V to redirect the slurry flow. The duct between
the absorber outlet and reheater inlet has experienced corrosion
to the extent that it has eaten through the ductwork. The
ductwork material is J" corten steel with a 60 to 80 mil
Ceilcote Flakeline 103 coating. After the coating either fell
off the wall or was eroded away, it was not long before corrosion
took its toll. This area was not the only area in which we
experienced problems with coatings. In the sump, the walls between
the abosrober were lined with Flakeline and after the scale was
removed a thorough inspection of the walls showed areas where the
lining was gone and corrosion was taking place. The other half
-------
of the sump, that area below the venturi, is lined with
1-J inches of Kaocrete refractory over the Flakeline and
appeared to be in excellent condition. The walls below
the absorber have since been cleaned and covered with
refractory. Severe corrosion of the reheater bundle support
frames has taken place. In some areas there is little,
if any, of the carbon steel remaining.
One of the absorber recirculation pumps was
damaged when a check valve in the pump discharge
piping failed and portions fell into the pump.
The pump impeller was damaged and its rubber lining
destroyed. From the debris found in the pump, it
appeared that the rubber lining had eroded on the check,
allowing the base material to corrode and ultimately fail.
B&W reviewed the need for check valves in the pump
discharges and recommend that they be removed. They have
since been removed and the spool pieces relined with rubber.
Unit #1 came down for a boiler, turbine and scrubber
overhaul in mid October for an outage that lasted more than
four months. This gave us an opportunity to conduct an
extensive inspection and maintenance program of all scrubber
components. Approximately $200,000 was spent during this
outage on maintenance material and labor in an attempt to
have the scrubber in like-new condition upon its return to
service.
Sludge Treatment
There has been little in the way of new development
at Will County regarding the sludge treatment and disposal.
The treatment methods remain the same as reported in previous
meetings, that is, thickener underflow or retrieved ponded
sludge is placed in concrete mixing trucks along with fly
ash and lime and then transported to a dumping site. Until
September 1975, the treated material was dumped into a
seven acre on-site clay lined disposal basin where the
material would set up after one to two weeks depending on the
ambient air conditions. Since September 1975 the Material
Service Company, a local quarry and concrete products firm,
has been transporting the treated sludge from our facilities
directly to a permanent disposal site which they operate.
228
-------
During 1975, Illinois EPA gave a local group a landfill
permit to use the stabilized sludge already in our on-site
disposal basin as fill for a future golf course. Over
eighty thousand tons of solidified treated sludge have been
removed from our basin and transported to the landfill site
at a cost to us of $2.35 a ton. There is still additional
material in our basin that needs to be removed when weather
permits.
Two years ago it was noticed that the solidified
treated sludge stockpiled in our on-site disposal basin
was crumbling at the surface. Investigation into this
phenomenon by Dr. Berger of the Civil Engineering Department,
University of Illinois has shown that the treated material
loses its ettringite and gypsum properties after undergoing
several freeze-thaw cycles. Analysis by x-ray diffraction
showed a greater deterioration of these properties as the
number of cycles increased. The exact cause of this
phenomenon is not totally understood; however, Dr. Berger
feels there may be a carbonation of material during the
freeze-thaw which lowers the pH, thereby causing
deterioration of the ettringite. Perhaps removal of a
larger portion of the water from the sludge and compaction
of the treated material will reduce the porosity enough
such that there is a reduced effect from freeze-thaw cycles.
A belt type vacuum filter has been purchased for
¥ill County and should be in service by mid-summer of this
year. Once this filter is in service we will have to change
our methods of handling which should reduce the cost of
treating the sludge. We also anticipate seeing some
improvement in the characteristics of the sludge after the
removal of more water. The one big disadvantage to this
installation will be the additional water we are going to
have in our system which will make it all the more
difficult to maintain a closed water loop.
Costs to Own and Operate
Approximate operating costs for the Will County
scrubber are shown in Figure 6. These costs are based on
actual expenses incurred for operation and maintenance
through the year 1975 and are considered approximate because
there are a few invoices still outstanding for 1975; however,
the numbers appearing in Figure 6 should be considered
indicative of the cost to own and operate the scrubber
'system.
229
-------
The fact that the scrubber did not have the same
availability as the boiler necessitated that the values
for the tons of coal consumed, BTU's and net kilowatt
hours used in the cost analysis be proportioned to the
quantity of flue gas treated by the scrubber when in
service.
Included in the cost to own and operate are the
following items: annual carrying charges on the invested
cost, operating labor, maintenance, limestone, auxiliary
pov/er, rehea.t steam and sludge treatment. Included in
the auxiliary power cost is the cost for actual power
consumed and the annualized cost of replacement kilowatts
for that which is lost to the system because of the scrubber.
Reheat steam costs are based on the cost of the fuel
required to generate the steam which is removed from the
boiler cycle.
The true meaning of maintenance costs is hard to
assess when taken over a short time interval. One would
not expect maintenance costs to be as high the first year
as would be expected after two or three years of operation.
Therefore it is essential that a good data base be
established over a period of time to get the true picture
of what maintenance on these systems actually cost. The
cost for maintenance as shown in Figure 6 would have been
higher had all the 1975 charges been available for
inclusion in the dollar figure shown. Not included in
the maintenance cost shown are the costs for the B-module
modification. The modification costs were investment
dollars rather than expense dollars because equipment of
a new design was installed. We anticipate higher
maintenance costs in the years ahead when greater operating
hours are attained on the system.
The sludge treatment cost includes labor, material
and hauling of the treated sludge to the off-site disposal
area. Also included is the maintenance cost of the sludge
treatment facilities. For the period January through November
27» 7^1 tons of dry solids scrubber sludge were treated
at a cost to Edison of $546,217. V/hen referring to dry
solids of scrubber sludge we are talking about the material
leaving the scrubber system with no additives or water
included. This would be the calcium sulfite and sulfate,
unreacted limestone and that fly ash removed in the scrubber's
venturi section. On this basis, it cost us $19.69 a ton (dry
scrubber solids) to treat and dispose of the material.
230
-------
To summarize, the cost to own and operate the
scrubber was about $22 a ton of coal, 1140 a million
BTU's or 13 mills per net kilowatt hour.
Future Expectations
While the main objective at Will County has been
to demonstrate scrubber operation on a full-sized unit, we
have, out of necessity, done a considerable amount of
experimentation in our efforts to make the system
operational. Three main reasons make it impractical to use
Will County for R&D projects that would require frequent
changes of hardware: (1). Will County Unit #1 while not
a base load unit, is an essential part of the Edison system
and has had a capacity factpr of 505' or greater over the
past years, (2). The large size means that any major
changes would require large amounts of time and money, and
(3). The tight physical configuration puts a constraint
on any changes that would require the addition or
re-arrangement of equipment.
It has become painfully apparent that operation on
high sulfur coal with an assured closed water loop will
require considerably more costly experimentation. From
past experience, just attempting to keep the system
operating will not provide solutions, if any, to the
problems at hand. We are in the process of developing a
test program to attempt to control, if this is possible,
the chemical cycle. The information gained from this
program, in addition to possibly improving operations at
Will County, may be of some value in designing, building
and operating the 450 MW scrubber slated for our
Powerton Station.
231
-------
TABLE I
1975 OPERATING HOURS & AVAILABILITY
Generating
Unit Operating
Hours
1975
Jan.
Feb.
Mar.
April
May
June
July
August
September
October*
November *
December *
6
* 21 week
684.5
666
609
638
744
642
689
564.5
720
194.5
0
0
,151-5
Scrubber
Operating
Hours
A B
676
662
604.5
252.5
628.5
389.5
0
0
0
0
0
0
3213
0
0
0
0
276
543
547
568
451.5
194
0
0
2579-
scheduled overhaul started
Scrubber
Availability
Based On Unit
Operating Hrs.
A B
98.8%
99.4%
99.3%
39.6%
84.5%
60.7%
0
0
0
0
0
0
5 51.8%
October 11,
0
0
0
0
37.1%
84.6%
79.3%
100.0%
62.7%
99.7%
0
0
41 . 9%
1975
Scrubber
Available
Hours
A . B
740
667
699
266.5
628.5
462
0
0
0
0
0
0
3^63
0
0
0
0
276
615.5
589.5
696
451.5
240.5
0
0
2969
Scrubber
Availability
Based on The Number
Of Hours In Each
Month
A B
99 . 5%
99.4%
94.0%
37.0%
84.5%
64.2%
0
0
0
0
0
0
39.5%
0
0
0
0
37.1%
85.5%
79.2%
93.5%
62.7%
32.3
0
0
33.9%
-------
FIGURE 1
SLUDGE
WASTE
POND
Q
THICKENER
PROPERTY PLAT
I38KV SWITCH YARD
TURBINE ROOM
BOILER ROOM
SERVICE
BUILDING
WASTE SLURRY AND RETURN
WATER PIPELINES
COAL
CONVEYOR
COAL
BREAKER
BUILDING
LIMESTONE
MILL
WET
SCRUBBER
BUILDING AND
CONVEYOR
RECLAIM
HOPPER
-------
RECLAIM
HOPPER
LIMESTONE
SILO
SYSTEM
FEEDER
_JCLASS1FIER
MILL
v
RECYCLE
TANK a ^
PUMPS
r
SLURRY
STORAGE
TANK
v
TO WET
SCRUBBER
-------
Recycle and
make-up water
Commonwealth Edison Company
Will County Station - Unit No. 1
APC-1
FIGURE 3
-------
FIGURE
BOILER
FLUE GAS PATH
STACK
EXISTING
ID FAN
BYPASS
DAMPER
ELECTRO. /
r-
\
NEW ID
'BOOSTER
FAN /
-VENTURI
REHEATER
' DEMISTER
SUMP
A3SORBER
-------
FIGURE 5
SLURRY RECiRCULATIOfel SYSTEM
TO SLUDGE
WASTE
POND
VENTUR!
PUMPS
VENTUR1 \
SUMP
VENTURI
REC1RCULATION
TANK
ABSORBER
ABSORBER
RECIRCULATiON
TANK
FROM MILL
SYSTEM
ABSORBER
PUMPS
-------
FIGURE 6
'WILL COUNTY UNIT 1 WET SCRUBBER
SCRUBBER SYSTEM
Carrying Charges
on $14,900,000
Labor (Operating
and Technical)
Maintenance (Labor
and Material)
Limestone
Auxiliary Power
Reheat Steam
APPROXIMATE
& ANNUAL COST
$2,280,000
.81,709
256,656
62,726
586,875
72,595
1975 COST TO
&/TON OF COAL
$ 12.54
0.45
1.41
0.54
3.23
0.40
OWN AND OPERATE
ei/MMBTU
65.3 0
2.3
7.4
1.8
16.8
2.1
MILLS/KWHR
7.47 Mills
0.27
0.84
0.21
1.92
0.24
$3,340,561
$ 18.37
95.7 0
10.95 Mills
SLUDGE TREATMENT
Carrying Charges
on $642,000
Sludge Treatment
98.000
546,217
0.54
3.00
2.8
15.7
0.32
1.79
$ 644,217
$ 3.54
18.5 0
2.11 Mills
SCRUBBER & SLUDGE TREATMENT TOTAL COST
$3,984,778 $ 21.91
114.2 0
13.06 Mills
Notes:
1. Scrubber system has a 14 year life.
2. Sludge treatment cost includes hauling to an off-site disposal area.
3. A-module on line 3213 hours, B-module on line 2580 hours. 181,886
tons of coal consumed, 3-49 X 10"!2 BTU and 305.1 X 10° net kilowatt
hours are proportioned to scrubber on line hours.
4. Coal average: 9604 BTU/lb, 1.596 sulfur.
5. Unit #1 12 month capacity factor for 1975 was 49.496.
12/30/75
238
-------
FIGURE 7
1O 11 12 13 14 15 16 17 18 19 ZO 21 22 23 24 25 26 27 28 29 3O 31
239
-------
n
UNIT
o
(0
la
CYCLONE, jn
110
A-MODULE SCRUBBER
B-MODULE SCRUBBER
.
'-fO.
NOZZLES.
IJN.IT.
A.
-IOUR
.
PI
•
-------
UNIT
A-MODULE SCRUBBER
B-MODULE SCRUBBER
-------
UNIT
A-MODULE SCRUBBER
B-MODULE SCRUBBER
-------
UNIT
V-MQDULE SCRUBBER
B-MODULE SCRUBBER
ICHAN
TO.P
REVENT. WATER. hV.ERBORAlJlNG.
1.11.HQIR5..
. TRJP-PAMPERS. .
UTESi
.
UNE)4:RSPR|AY. NOZZLES.
5 HOtJRS
.
.
•
Mm
!UNDERGO[NG
iiiiPt-:
m
REffTjRN
'1, H
-------
VJ1
'
I
-------
UNIT
! *.
Li
. cyciQNt;
"I
HOURS
>DULE SCRUBBER
B-MODULE SCRUBBER
, .DEMIST]'5 ft.
i-LS.
. REfLACEMEMl
L.J
o
86 ,H
. i
STEW jb&
1 .OUR
A
-------
A-MODULE SCRUBBER
B-MODULE SCRUBBER
IT
NO. .UNIT.
CYCLONE.
-------
UNIT
A-MODULE SCRUBBER
B-MODULE SCRUBBER
MISTER &
-R. t
IPLAOEMEN'T
I
SS
.
DEHiCSEH
198 HOURS
BALA:
:
, | CLjEAN
y3SQR_BES
0
HQURS
T
'IMP
£CJRJEN
-------
UNIT
A-MODULE SCRUBBER
B-MODULE SCRUBBER
DlMIStEB.&t.
UKUWXX-.3OIM
R] HEATER. CDII
CHEDULED. JVERIJAUL
rnr t r%*r%^*ii»t*i n
E2ULEI!
JlAUL
-------
MHI FLUE GAS DESULFURIZATION SYSTEMS APPLIED TO
SEVERAL EMISSION SOURCES
M. Hirai, M. Atsukawa, A. Tatani, and K. Kondo
Mitsubishi Heavy Industries, Ltd.
Tokyo, Japan
ABSTRACT
Since the development of its lime/limestone-gypsum recovery
process, Mitsubishi Heavy Industries, Ltd. (MHI) sold thirty-two
(32) units of MHI FGD Systems in Japan. Twenty (20) units are operating
smoothly, eight (8) are under construction and four (4) are committed
by a contract. The total capacity of these installations amount to
11,282,000 scfm, treating gases from boilers, sintering plants and a
copper smelter.
This paper describes the features of MHI FGD System, operating
experience of its commercial scale units and its applicability to coal-
fired boilers.
Further, MHI recently developed a process of removing SO and
NO simultaneously by lime/limestone scrubbing, and its 1,200 scfm
pilot test was completed in 1975. Its outline is also reported
in this paper.
249
-------
1. INTRODUCTION
The development of MHTs lime/lime stone-gypsum recovery
process started 15 years ago, and its intermediate progress report
was made public at the Second International Lime/Limestone Wet
Scrubbing symposium, back in 1971. (Reference 1)
At that time construction was in progress for a large prototype
unit with the capacity to treat 20% of the flue gas from a 156 MW
power unit. This prototype plant started up in March, 1972 and as
its operation was "very satisfactory (Reference 2), MHI has been suc-
cessful in marketing its systems in Japan. When limited to utilities,
MHI's share in the.market is about 56% in terms of treated gas flow.
(Reference 3)
MHI's system development is shown in Figure 1, and its delivery
record is listed in Table 1.
Among the 20 units in operation, 15 units are for oil-fired
boilers, 4 units are for sintering plants (steel mills) and the remaining
one unit is for a copper smelter. The capacity of most of these in-
stallations ranges from 250,000 to 740,000 scfm in terms of gas volume
treated, and among the units so far installed, the largest capacity of
one module of a scrubber is 560,000 scfm. SOj, concentration in the
flue gas ranges from 500 to 23,000 ppm. Lime or Limestone is used
as absobent and high quality gypsum is recovered as a by-product
which is used for cement and wallboard production.
The prototype plant mentioned above has so far been operated
for about 4 years, and it is now being operated as a 156 MW full-
scale plant, in combination with a plant which treats the remaining
80% of the flue gas. Further, another 156 MW full-scale plant is
under construction in the same power station.
In evaluating a lime/limestone scrubbing system, the following
three items are considered to be important.
1) Reliability
2) Economics
3) Sludge disposal
From the start point, MHI's aim were to fulfill the above re-
quirements and so far have been quite successful.
250
-------
This paper describes the features of the MHI Lime/Limestone -
Gypsum Process, operating experience of its commercial scale system
and its applicability to coal-fired boilers.
Recently, NOX removal has become a serious problem. For
several years, MHI has been conducting a series of R & D programs
on both wet and dry NOX removal processes. One of the successful
developments is a process of removing NOX and SOX simultaneously
in a lime/limestone scrubber at high efficiency, by use of a water
soluble catalyst. As its pilot test was completed in 1975, its outline
is introduced in this paper.
251
-------
2. SYSTEM DESCRIPTION
The MHI FGD System is a wet non-regenerative system which
uses lime or limestone as absorbent and recovers high quality gypsum
as by-product. A simplified flow scheme of the system is shown in
Figure 2.
The flue gas is fed into the precooler to be cooled down, most
of its particulates being removed there at the same time before enter-
ing the SC>2 scrubber. The absorbent slurry for SC>2 absorption, and
the gypsum "seed" slurry for scale prevention, are fed to the scrub-
ber. The cleaned flue gas passes through the mist eliminator and is
then reheated to restore buoyancy.
The scrubber slurry flows into the sulfite oxidizer where all
the calcium sulfite in the slurry is oxidized by air, yielding gypsum
slurry. A portion of the gypsum slurry is recycled to the scrubber
and the remainder is transferred to the gypsum recovery section of
the system. Here, the gypsum slurry is concentrated by a thickener
and the underflow slurry is fed to centrifuges, yielding by-product
gypsum. Most of the thickener overflow is reused for absorbent
preparation. Slowdown water from the precooler and part of the
thickener overflow are fed to the water treatment section of the
system.
Each constituent of the MHI FGD System has the following ob-
jectives and features.
2. 1 Precooler
Main objectives are as follows:
l) Removal of dust from the incoming gas to obtain high quality
gypsum.
2) Cooling of the flue gas to protect plastic materials used in the
SC>2 scrubber.
3) Humidification of the flue gas to avoid any local evaporation
which may lead to deposit formation in the SC>2 scrubber.
A spray tower is used to economically achieve the above require-
ments. The spray pump capacity in terms of L/G is normally 14
and gas stream pressure drop (including demisters) is about 1 1/2
in.
252
-------
2. 2 SO2 Scrubber
Important requirements of this equipment are as follows:
l) Stable and high SC>2 removal efficiency even at severe load
fluctuations.
2) Scale-free operation.
3) Low pressure drop (i. e. low power requirement).
4) Low mist entrainment at the outlet.
In order to accomplish .the above requirements, MHI has developed
a packed spray tower employing low density "grid" packings.
Simple structure of the "grid" packings enables the scrubbing slurry
to be distributed uniformly on its surface, thus, giving a large ab-
sorption area. Another feature of the "grid" packed tower is its
low mist entrainment at the outlet which contributes to the reliability
of the mist eliminator. The scrubbing slurry pumping rate in terms
of L/G is 50 to 140.
Scale-free operation is accomplished by using relatively high L/G
and adding gypsum "seed" slurry to the scrubber. The gas stream
pressure drop is about 1 1/2 in.
2. 3 Mist Eliminator
The mist must be removed from the gas at high efficiency without
scaling of the eliminator surfaces. All of the MHI FGD Systems
that are now in operation are using vertical chevrons, giving very
good service. Pressure drop is within the range of 1/2 to 1 in.
2.4 Reheater
Reheating the scrubber gas is necessary to restore buoyancy, to
reduce visible steam plume, and to protect downstream equipment
from possible corrosion.
2.5 Absorbent Preparation
This section of the process is important bacause it gives great
influence on the following.
1) SC>2 removal efficiency and absorbent utilization.
253
-------
2) By-product gypsum quality.
In order to achieve this, the ultimate method is to make the particle
size of the absorbent slurry as small as possible.
2. 6 Sulfite Oxidizer
The oxidation reaction of calcium sulfite to calcium sulfate depends
on dissolution and diffusion of oxygen into the slurry. MHI is using
a rotary, atomizer, an air atomizer of unique structure, which gene-
rates fine air bubbles at operating pressure of about 50 psig. The
oxidation rate is also improved by lowering the slurry pH.. A small
amount of sulfuric acid is fed to the oxidizer for this purpose.
2. 7 Gypsum Recovery
Gypsum crystals produced in the oxidizer are relatively large, blocky
crystals that settle easily. Therefore, the gypsum slurry can be con-
centrated in a reasonable size thickener and dehydrated by conventional
centrifuges. The by-product gypsum contains only 5 to 10% of free
water. Gypsum is a stable, harmless compound that can be used for
wallboard production and as a retarder for cement setting. In the
case of over-supply, the by-product gypsum is more suitable for land-
fill disposal than sulfite sludge.
2. 8 Water Treatment
Some water is disposed of to keep the concentration of chloride and
other impurities in the system ar a safe level.
Bleed streams from the precooler and the thickener overflow are
treated to meet the local water pollution regulations.
2. 9 System Control
The controls for the MHI FGD System depend upon the type of ab-
sorbent used. In lime scrubbing, the slaked lime slurry feed is
controlled by the pH of the recirculating scrubber slurry. In lime-
stone scrubbing, the limestone slurry feed is controlled by the SO2
in the flue gas and the pH of the recirculating scrubber slurry.
254
-------
3. OPERATING EXPERIENCE
MHI has experience in construction and operation of the FGD
systems applied to several emission sources. They are oil-fired
boilers, sintering plants (steel mills) and a copper smelter.
Each emission source has different flue gas conditions, and
consequently, the associated FGD system runs somewhat differently
from one another. Operating data obtained from the commercial plant
operation are described in Table 2.
Operating experiences in the commercial plants, problems en-
countered, and solutions to these problems are outlined below.
3.1 Precooler Operation
Problems on materials of construction were experienced due to low-pH
(1 to 1.5) and corrosive nature of the recirculating spray water.
1) Pin-holes on rubber linings due to bad -workmanship caused corro-
sion on carbon steel underneath.
This problem was overcome by setting stringent standards for
lining works and pin-hole inspection.
2) In an installation for a sintering plant, the incoming gas laden
with oily matter caused swelling of rubber linings.
This problem was solved by using oil resistant rubber linings.
3) In an installation which used sea water to prehumidify the flue
gas, accumulated chloride ions in the recirculating spray water
caused corrosion on stainless steel valves. The problem was
solved by using rubber lined valves.
3.2 Scrubber Operation
3.2.1 SO, removal efficiency and absorbent utilization
£*
Independent of flue gas conditions and types of absorbent used, the
concentration at scrubber outlet ranges from 10 to 60 ppm, or more
than 90% in terms of removal efficiency.
255
-------
Past operating data shows that the absorbent utilization ranges from
90 to 99% or on the average of 95%. One of the reasons why the
absorbent utilization is high even in limestone scrubbing is that most
of the limestone in Japan is of high calcium in nature.
3.2.2 Scrubber control
Using the pH of the recirculating scrubber slurry as a control factor,
all of the scrubbers are running with excellent stability. For example,
the SO concentration of flue gases from sintering plants changes from
800 to 1,200 ppm every 20 minutes. Even at this kind of fluctuation,
the FGD system operated with high and stable SO2 removal.
3.2.3 Scrubber scaling
One of the important requirements in slurry scrubbing is scale-free
operation of scrubbers. In the early stage of development, some
scaling problems were experienced due to poor operating procedures.
According to MHI's experience, the following considerations are vital
in preventing scaling.
1) Addition of gypsum "seed" crystals to provide seed sites for
calcium sulfate crystallization.
2) Application of relatively high L./G and selection of a scrubber
with simple structure, to maintain chemical conditions of the
scrubbing slurry uniform, and to keep all of the internals wet
and clean.
The above method for preventing scaling was proved to be a good way,
even in units for sintering plants where the sulfite oxidation in the
scrubber reaches as high as 100%, due to high QZ concentration of
the incoming flue gas.
3.3 Mist Separation
The scrubber outlet mist loading ranges from 4 to 8 gr/acf.
Using vertical chevrons, more than 99% of the entrained mist is
separated from the gas. Fresh water and a part of the scrubbing
slurry is used for washing the eliminator surfaces. Continuous
washing and intense intermittent washing were experienced, both
giving excellent service.
256
-------
3.4 Gas Reheating
Most of the installations reheat the scrubber gas directly by burning
low-sulfur fuel oil. Installations for sintering plants burn combus-
tible flue gases from steel mills.
One installation for small industrial boilers do not reheat at all.
The reheat temperature is rather high in Japan, usually in the range
of 250 to 285 °F.
3.5 Absorbent Preparation
Both for lime and limestone, recycle water is used for preparation of
tiie absorbent slurry. In lime slaking, the grits are ground in a wet
mill and carried along with the slaked lime into the scrubber. Lime-
stone is ground in a wet mill, but recent installations are using fine
limestone powder.
3.6 Lime vs. Limestone
Recent installations tend to use limestone rather than lime. According
to MHI's experience, the difference between limestone and lime can be
outlined as follows.
1) Lime is more reactive, and therefore lime scrubbing is much
easier to control.
2) Sulfite oxidation rate is much higher in limestone scrubbing.
3) Absorbent preparation is much easier with limestone powder.
In lime slaking, the substitution reaction between lime and the
dissolved magnesium sulfate in recycled water may cause calcium
sulfate scaling. Here again,, the addition of gypsum "seed"
crystals in the slaker is very effective for scale prevention.
4) Limestone is more economical, safer and easier to handle.
3.7 System Reliability
All of the installations are running with high availability.
Availability data on some installations are shown on Table 3.
257
-------
4. APPLICATION ON COAL-FIRED BOILERS
Since most of the utilities in Japan are using oil in their boilers,
all of the MHI FGD System installations for boilers are treating oil-
fired gases. If an electrostatic precipitator is installed before the
system, main differences between oil and coal-burnt gases would be
in compositions of dust and gas. MHI has been conducting field pilot
tests on a coal-fired boiler in Japan. Test results obtained up to
now indicates that the MHI gypsum recovery process is applicable
on coal-fired boilers without problem. On the other hand, applica-
bility of the system to non-recovery (throwaway) processes has been
tested also. If a non-recovery process is favorable, which is the
case in the United States (Reference 4), the MHI FGD System can
meet this requirement, and can be simplified as follows.
1) Precooler can be eliminated.
2) Absorbent preparation section can be simplified.
3) Gypsum recovery section can be eliminated.
4) Sulfite oxidizer capacity is minimized to gypsum "seed" pro-
duction.
For application in the United States, tests on dolomitic lime/
limestone scrubbing are also in progress. Through these tests,
influences of chloride and other impurities on the system have been
investigated. Test results up to now indicates that the system is
quite feasible. Moreover, studies on materials of construction,
process equipment, process economics are being continued for further
development of the system.
258
-------
5. MHI SIMULTANEOUS NOV/SC> REMOVAL PROCESS
•"• X
MHI's research for the NOX removal process has been aimed
at the development of reasonable and economical systems which can
treat flue gases ranging from clean flue gases of LNG-firing boilers
to dirty flue gases of high sulfur fuel-firing boilers, sintering plants
(steel mills), etc. MHI has developed both dry and wet processes of
NOx removal and has already obtained orders for the NOX removal
units. For dirty flue gases requiring both NOX and SOX removal,
MHI is considering that the use of the wet process is most reasonable
for them, in view of its capability of simultaneous NOX/SOX" removal
in a single scrubber and from the standpoints of economics, operation
and maintenance, installation area, etc.
Recently, MHI has successfully finished its preliminary stage
of development for a simultaneous NOX/SO removal process. By
adding an 03 generator and an ammonia recovery section to the MHI
FGD System, NOX and SO in the flue gas are removed simultane-
ously in a single scrubber, at high efficiencies. The scrubber is
identical to those used in the MHI FGD System, lime or limestone
can be used as absorbent, and by-products are gypsum and ammonia.
5.1 Process Development
After a bench scale test (125 scfm capacity) for about a year,
a pilot plant test (1,250 scfm capacity) was conducted from July, 1975.
The flue gas from an oil-fired boiler contained 180 to 220 ppm NO^
800 to 1,100 ppm SOX, and 0.04 to 0.05 gr/scf dust. The obtained
efficiencies were 80 to 90% NOX removal and over 95% SOX removal.
Since the result was satisfactory, a prototype scale unit is planned
in the near future.
5.2 Process Description
A simplified flow scheme of the process is shown in Figure 3.
The flue gas is first cooled and dust collected in the precooler
and after the oxidation of NO to NO2 -by means of adding stoichio-
metric amount of O,, NO_ and SO_ are removed simultaneously in
the SO_/NO scrubber. The absorbent slurry, either slaked lime
slurry or limestone slurry, containing water soluble catalyst is fed
to the scrubber.
259
-------
Since the scrubber outlet slurry contains the N-S compounds
(defined as a mixture of nitrogen and sulfur base compounds) as
dissolved components, a portion of filtrate from the gypsum re-
covery section is sent to the ammonia recovery section where it is
decomposed to yield gypsum and ammonia.
Description of the remaining section of the process is identical
to those of the MHI FGD System.
5.3 Features
The outstanding features of the process can be summarized
as follows:
1) Most of the main equipment are exactly the same as that of the
MHI FGD System and their reliabilities are already known to be
high through actual operation.
2) Since NO and SO are removed simultaneously in one system,
the process has advantages in installation area, operation and
maintenance, economics, etc.
3) As the reductive property of SO, is used in absorption, the
process does not require any NO reducing agent. (In the
dry process, ammonia is added as the reducing agent)
4) The process is a recovery system which produces gypsum and
ammonia as by-product.
5) The process can be applied to the existing MHI FGD System by
simple reconstruction.
260
-------
FEASIBILITY TESTS
FOR OIL-FIRED BOILER
3,300 scfm PILOT UNIT
DATA FEED BACK FOR IMPROVEMENT
A
COMMERCIAL SCALE UNIT
FOR COPPER SMELTER
t\
COMMERCIAL SCALE UNITS
FOR SINTERING PLANTS
IT
IT
HIGH S02 SCRUBBING TESTS
TESTS ON THE INFLUENCE OF IMPURITIES
APCS RESEARCH CENTER
PERMANENT LAB. UNITS
1,200 scfm ,125 scfm
(HIROSHIMA TECHNICAL INSTITUTE)
SCALE-UP TESTS
FOR OIL-FIRED BOILER
62,000 scfm
PROTOTYPE UNIT
FEASIBILITY TESTS
FOR COAL-FIRED BOILER
1,200scfm FIELD PILOT UNIT
COMMERCIAL SCALE UNITS
FOR OIL-FIRED BOILERS
Figure 1. MHI FGD System Development
COMMERCIAL SCALE SYSTEM
DEVELOPMENT COMPLETED
FOR COAL-FIRED BOILER
( THROW-AWAY AND
V GYPSUM RECOVERY SYSTEMS
-------
FLUE GAS FROM
EMISSION SOURCE
FLUE GAS TO
CHIMNEY
MIST
ELIMINATOR
S02
SCRUBBER
PRECOOLER
REHEATER
LIME/
LIMESTONE
GYPSUM
RECOVERY
BY-PRODUC
WATER
TREATMENT
ABSORBENT
PREPARATION
OXIDIZER
SLUDGE) (PURGE
GYPSUM/
Figure 2. MH! FGD System CMHI Lime/Limestone-Gypsum Recovery Process)
-------
FLUE GAS FROM
EMISSION SOURCE
OT
GENERATOR
c*r\ 1
PRECOOLER
S02/NO?
SCRUBBER
WATER
TREATMENT
MIST
ELIMINATOR
ABSORBENT
PREPARATION
GYPSUM
RECOVERY
AMMONIA
RECOVERY
IF
D
D
D
D
LIME/
LIMESTONE
Figures. MHI Simultaneous N0x/50x Removal Process
I CD CD CD
D
D
a
D
FLUE GAS TO
CHIMNEY
BY-PRODUCT
GYPSUM
f-PRODUCl
.AMMONIA.
o
-------
Table 1. DELIVERY RECORD OF MHI FGD SYSTEM
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Owner
Kansai Electric Power Co.
Onahama Refining Co.
Yoshino Gypsum Co.
Kawasaki Steel Corp.
Kanaai Electric Power Co.
Tokyo Electric Power Co.
Tohoku Electric Power Co.
Kyushu Electric Power Co.
Kawasaki Steel Corp.
Kansai Electric Power Co.
Teijin Limited
Kawasaki Steel Corp.
Kawasaki Steel Corp.
Kuraray Co.
Mizushima Joint Thermal
Location
Amagaaaki
Onahama
Tokyo
Chiba
Kainan
Yokosuka
Hachinohe
Kanda
Mizushima
Amagasaki
Ehime
Mizushima
Chiba
Saijo
Mizushima
Gas Source
156 MW Unit
Copper Smelter
Industrial Boiler
Sintering Plant
600 MW Unit
265 MW Unit
250 MW Unit
375 MW Unit
Sintering Plant
156 MW Unit
Industrial Boiler
Sintering Plant
Sintering Plant
Industrial Boiler
156 MW Unit
Capacity,
scfm
62,000
57,000
8,000
74,000
248,000
248,000
235,000
341,000
465,000
232,000
167, 000
558,000
260,000
131,000
379,000
/IB
Absorbent
Lime
Lime
Slaked Lime
Lime
Lime
Limestone
Lime
Lime
Lime
Lime
Lime
Lime
Lime
Lime
Lime
UJ. J «II1. , 1 /
Start-Up
1972
1972
1973
1973
1974
1974
1974
1974
1974
1975
1975
1975
1975
1975
1975
Power Co.
16.
Toyobo Co.
Iwakuni
Industrial Boiler
124,000
Lime
1975
-------
17.
18.
19.
20.
21.
22.
23.
24.
t-o
a 25.
26.
Niigata
Power
Chubu
Chubu
Kyushu
Kyushu
Kyushu
Kyushu
Tohoku
Owner
Joint Thermal
Co.
Electric
Electric
Electric
Electric
Electric
Electric
Electric
Kawasaki Steel
Power Co.
Power Co.
Power Co.
Power Co.
Power Co.
Power Co.
Power Co.
Corp.
Kashima-minami Joint
Thermal Power
27.
28.
Kansai
Niigata
Electric
Co.
Power Co.
Joint Thermal
Location
Niigata
Owase
Owase
Karatsu
Karatsu
Ainoura
Ainoura
Niigata
Chiba
Kashima
Amagasaki
Niigata
Gas Source
350
375
375
500
375
375
500
600
MW
MW
MW
MW
MW
MW
MW
MW
Sintering
160
156
350
MW
MW
MW
Unit
Unit
Unit
Unit
Unit
Unit
Unit
Unit
Plant
Unit
Unit
Unit
Capacity,
scfrn
328,
744,
744,
353,
452,
452,
452,
260,
446,
267,
294,
328,
000
000
000
000
000
000
000
000
000
000
000
000
Km
Absorbent
Limestone
Lime
Lime
Limestone
Limestone
Limestone
Limestone
Limestone
Lime
Limestone
Lime
Limestone
o! Jan. . — WT-o
Start-TJp
1975
1976
1976
1976
1976
1976
1976
1976
1976
1976
1976
1977
Power Co.
29. Sakata Joint Thermal
Power Co.
30. Sakata Joint Thermal
Power Co.
31. Kawasaki Steel Corp.
32. Chugoku Electric Power Co.
Sakata
Sakata
350 MW Unit
350 MW Unit
Mizushima Sintering Plant
Shimonoseki 400 MW Unit
682,000 Limestone
682,000 Limestone
465,000 Lime
744,000 Limestone
1977
1977
1977
1977
-------
Table 2. OPERATING DATA OF MHI FGD SYSTEM
ITEM """ """ — — SOURCE
Inlet S02 ppm (Volume)
SO2 Removal %
Absorbent Utilization %
in Scrubber
Scrubber Control pH
Oxidation in Scrubber %
Slurry Concentration % (Weight)
Inlet Gas Composition
O2 % (Volume)
CO2 % (Volume)
H2O % (Volume)
Dust Grains /scf
Remarks
Oil-Fired Boilers
500 - 1,500
90 - 99
90 - 99
/6.5 - 7.0 (Lime)
M.O - 5.8 (Limestone)
30 - 100
10 - 14
(Pre -cooler Inlet)
3-6
9 - 12
9-12
0.0085 - 0.0175
(ESP, Outlet)
0.04 - 0.08
(Boiler A/H Outlet
One Scrubber System
, 12 Units
Two Scrubber System
1 Unit
Sintering Plants
(Steel Mills)
600 - 1,200
90 - 97
95 - 99
6.5 .- 7.5
70 - 100
10 - 14
(Pre-cooler Inlet)
12 - 17
4-8
6-12
0.0175 - 0.04
(ESP Outlet)
Inlet Cl : 20 -SOppnr
(as HC1)
Inlet Oily Matter:
-0.025 gr/scf
Dust Components:
Fe,Mn, Si, Pb,K,Na,
Ca, Mg, Al, Zn, Cu, etc.
Copper Smelter
20,000 - 29,000
90 +
- 99
,3-3.5 (T)
^9 - 10.5 ©
0-20
10 - 13
(Scrubber Inlet)
6
10
23
0.06
Two Scrubber System
(I) No. 1 Scrubber
© No. 2 Scrubber
i
-------
Table 3. AVAILABILITY OF MHI FOP SYSTEM
As of April. 1975
Owner
(Location)
1. Kansai Electric Power
Co. (Amagasaki)
2. Onahama Refining Co.
(Onahama)
3. Yoshino Gypsum Co.
(Tokyo)
4. Kawasaki Steel Corp.
(Chiba)
5. Kansai Electric Power
Co. (Kainan)
6. Tokyo Electric Power
Co. (Yokosuka)
7. Tohoku Electric Power
Co. (Hachinohe)
8. Kyushu Electric Power
Co. (Kanda)
Start-Up
(exclusive
of test run
Mar., 1972
Nov. , 1972
Aug., 1973
Nov. , 1973
Dec., 1973
Apr., 1974
May. , 1974
Dec., 1974
Emission
Source
Operating
Mrs (A)
16,000
24, 120
12,048
11,520
8,760
8,730
4,800
2,880
FGD
System
Operating
Mrs (B)
14, 600
23,780
11,328
11,456
8,760
7,886
4,080
2,880
Availability
Q3)/tA) x 100
%
91.5
98. 6
94.0
99.4
100.0
90. 3
85.0
100.0
Remarks
(Capacity: scfm x SO£ ppm)
Boiler stop t Apr. "74-Jan. '75
(62,000 x 800)
(57,000 x 23,000)
(8,000 x 1,500)
(74,000 x 500)
Boiler stop : Dec. '74-Mar. '75
(248, 000 x 300)
Boiler stop '. May '74
(235,000 x 800)
Boiler stop : Dec. '74
(235,000 x 800)
(341,000 x 500)
I J
0-.
-------
References
1. Uno, T., M. Atsukawa, K. Muramatsu, 834 - 849, Proceedings
of Second International Lime/Lime stone Wet-Scrubbing Symposium,
New Orleans, 1971.
2. Atsukawa, M., K. Matsumoto, N. Shinoda, M. Kirnishima,
Y. Ouchi, "Removal of SO_ from Flue Gases by Wet Processes",
M.H.I. Technical Review, February, 1974.
3. The data extracted from "Heavy Industries News", published in
Japan, December, 1975.
4. Slack, A.V., "Scrubber survey: a lime /lime stone trend",
Electrical World, October 1, 1974.
Figures
1. MHI FGD System Development
2. MHI FGD System (MHI Lime/Lime stone-Gyp sum Recovery Process)
3. MHI Simultaneous NO /SO Removal Process
x x
Tables
1. DELIVERY RECORD OF MHI FGD SYSTEM
2. OPERATING DATA OF MHI FGD SYSTEM
3. AVAILABILITY OF MHI FGD SYSTEM
268
-------
STATUS OF FLUE GAS DESULFURIZATION
USING ALKALINE FLY ASH FROM WESTERN COALS
Harvey M. Ness, Research Chemist
Everett A. Sondreal, Research Supervisor
Philip H. Tufte, Chemical Engineer
U.S. Energy Research and Development Administration
Grand Forks Energy Research Center
Grand Forks, North Dakota
ABSTRACT
It is the purpose of this paper to report on the status of flue
gas desulfurization utilizing the alkali in fly ashes derived from
Western coals. Most of the utility scrubbers operating in the Western
United States have been installed for particulate removal. These
particulate scrubbers, operating without an added reagent such as
lime or limestone, have been found to remove appreciable amounts of
sulfur dioxide from the flue gas and to experience sulfate scaling.
Sulfur dioxide removal occurs because of the inherent alkalinity
of the fly ashes produced from Western coals. The deliberate use of
fly ash alkalinity for flue gas desulfurization has been investigated
in four separate pilot-scale studies; two commercial-scale installations
have been built and a third is being constructed. The properties of
Western coals, their sulfur dioxide emissions, and the design and
performance of selected commercial and pilot scrubbers are discussed
in this paper. An appendix lists data on operating scrubbers.
269
-------
STATUS OF FLUE GAS DESULFURIZATION
USING ALKALINE FLY ASH FROM WESTERN COALS
INTRODUCTION
The Western reserve base for measured and indicated coal in place, as
defined by the U.S. Bureau of Mines, totals 2l6 billion tons1. Future plans for
mine expansion will increase capacity in the West above 200 million tons per
annum by 19832. This figure does not include tentative plans for a number of
coal gasification plants. Considering these projections, stack gas cleaning
technology for burning Western coals will assume much greater importance in the
future than at present.
The Western coal reserves include lignite, subbituminous, and bituminous
coal, with the lower rank coals predominating. An important property of almos^
all Western coals is that they contain far less sulfur than the 2 to 3 pet and
higher typical of Eastern and Central coals. Unfortunately, sulfur content in
Western coals (averaging 0.7 pet) is not generally low enough to meet new source
emission standards. Typically, an average sulfur dioxide removal of 30 to ^0
pet is required to meet the Federal standard of 1.2 lb S02/MM Btu, and higher
removals are required to meet more stringent State and local standards. Since
sulfur oxide emission standards are based on heat released, variations in
heating value according to rank have an important effect on the coal sulfur
content that is equivalent to the emission standard, as shown in Table 1.
TABLE 1 COAL SULFUR CONTENT EQUAL TO EMISSION STANDARDS
Coal
North Dakota
lignite
Higher
heating
value ,
Btu/lb
6,800
Coal sulfur
equal to the
Federal
standard of
1.2 lb S02/MM
Btu, pet
O.Ul
Montana
subbituminous 8,600 .52
Arizona
(Black Mesa)
bituminous 11,000 .66
While the 0.7 pet average sulfur content in Western coals does not
satisfy the Federal standard, it does make stack gas cleaning potentially easier
to achieve and less costly than for high-sulfur Eastern coals, provided that
emission standards are not raised to cancel out the advantage. In addition,
the alkaline nature of Western fly ash provides an opportunity for design
innovations in flue gas desulfurization.
270
-------
Ash content in Western coals can vary greatly, with the U to 20 pet shown
in Table 2, representative of the overall range. The ash content varies
significantly between mines, and even between locations within mines. The
quantity of fly ash in stack gas depends on boiler design as well as on coal ash
content. For a pulverized coal-fired boiler, fly ash leaving the boiler represents
approximately 80 pet of the coal ash. Resulting particulate loadings are
typically 2 to 10 gr/scfd. For a cyclone-fired boiler, fly ash leaving the
boiler is approximately ho pet of the coal ash. The corresponding particulate
loadings are 1 to 5 gr/scfd.
A very important characteristic of many Western coal ashes is their high
content of sodium oxide, magnesium oxide, and particularly calcium oxide (see
Table 2). The alkali content tends to be highest in lignite, the lowest rank
coal, and progressively less prevalent in the subbituminous and bituminous
coals. Variations in alkali content are also influenced by the minerology of
the overburden and the course of ground water movement. Alkali content in
Western coal ash varies from under 10 to over 50 pet, with important variations
occurring within individual mines.
A guideline in assessing the importance of the alkali in Western coal is
the ratio of the alkali to coal sulfur. For a coal containing 7-5 pet ash and
20 pet alkali in the ash, the total alkali is chemically equivalent to slightly
more than 120 pet of a 0.7 pet sulfur content. For some lignites, the alkali/sulfur
ratio can be several hundred percent. Thus there is ample alkali to interact
importantly with sulfur oxides in a wet scrubber in burning Western coals.
Required Removal Efficiencies
Required removal efficiencies are determined by coal sulfur content, sulfur
retention on ash during combustion, and the emission limits established by law.
As stated previously, the average 0.7 pet sulfur content of Western coals does
not permit such coal to be burned without flue gas desulfurization under the
Federal emission standard of 1.2 Ib S02/MM Btu. Retention of sulfur oxides on
ash during combustion may lower the sulfur dioxide emission by 10 to ho pet for
lignites , but this effect does not guarantee compliance with the Federal
standard. Considering more stringent state and local standards, there are no
coals that will meet the standards of Clark County, Nevada (0.15 Ib/MM Btu), New
Mexico (0.3h Ib/MM Btu), Nevada (O.hO Ib/MM Btu), or of Colorado after 1980
(0.35 Ib/MM Btu); it is doubtful, that any would reliably meet the Arizona
statewide standards (0.8 Ib/MM Btu). The removal efficiencies required to meet
the more stringent standards are shown in Figure 1 for a Western subbituminous
coal. At the 0.7 pet coal sulfur level, the required sulfur dioxide removal is
increased from about 30 pet to meet the Federal standard to 90 pet to meet the
Clark County, Nevada standard.
Capital and operating costs for stack gas cleaning must be expected to rise
steeply as high percentage removals are required for substances that are
initially present at low concentrations. For an approximate estimate based on
principles of engineering design, it can be assumed that equipment size and
power requirements for wet scrubbing will increase in proportion to the logarithm
of one over the required exit concentration (size and power a log i exit), and
271
-------
100
c
o.
O
2
Ul
a:
a
UJ
oc
ID
o
UJ
ac
80 -
60 -
40 -
20 -
Subbituminous coal
8,600 btu/lb
500 mw
.4% sulfur in coal
Clark County New
Nevada Mexico Nevada
.2 .4 .6 .8 1.0 1.2 1.4
S02 EMISSION STANDARD, Ib/mm btu
1.6
1.8
Figure I. - Removal efficiencies required for stack gas cleaning of Western coai.
-------
TABLE 2 SELECTED ANALYSES OF ASH IN WESTERN COALS'
Coal
State
Sample avg. . . .
Lignite
North Dakota
212
Subb ituminous
Montana
125
Wyoming
Big Horn
12
Arizona
Black Mesa
1
New Mexico
Nava j o
2
' Bituminous
New Mexico
McKinley
1
Colorado
Hawks Nest
3
Ash, percent of coal
6.2 9.3 1+.8 7.5
Oxide constituents, percent of ash
20.2
8.0
SiOp
AloOo
FeoO-a
TiOo
PoOs
CaO
Ma;0
Na20.
K20
303
19.7
11 . 1
9.1
.1+
.3
2k. 6
6.9
6.5
.k
19-5
35.5
18.7
7.8
.7
.3
15.6
1+.1+
1.7
.1+
13.1+
27. k
12.7
13.9
.6
.5
16.6
5.5
2.2
.5
17.0
1+2.0
18.1
5.7
.8
.6
17.8
2.1+
1.1+
.3
8.2
55.6
26.2
6.1
.6
.5
3.9
.8
1.5
.6
3.2
5H.7
21.6
7.0
1.0
.0
6.5
1.2
1.6
.8
5.8
1+1+.8
28.3
11.5
.8
.7
5.6
1.9
.6
.5
l+.O
-------
capital cost will increase as the 0.6 pover of size. Under these assumptions,
equipment size would triple between 50 pet and 90 pet removal, and would double
again at 99 pet removal. Capital cost would first double and then rise by a
further 50 pet for the same increases in removal. These figures are hypothetical
and they make no allowance for improved design or use of more effective reagents.
However, they are close enough to reality to validly demonstrate that high costs
will have to be paid to achieve improved control of stack gas emissions.
The following sections provide up-to-date information on the capital
investment and operating experiences of the Western coal-burning utilites
experiencing sulfur dioxide removal in particulate scrubbers. It also provides
up-dated information on utility operated scrubbers that are deliberately utilizing
fly ash alkali for sulfur dioxide removal, and also, on current research results
on fly ash alkali utilization and scaling problems.
Arizona Public Service Company, Four Corners Plant^'
The Four Corners Plant at Farmington, NM has three pc-fired boilers (two
175 MW, and one 225 MW) that are equipped with Chemico venturi scrubbers for
particulate removal, with two scrubber modules on each boiler. Two additional
boilers of 755 MW each are equipped with electrostatic precipitators (ESP's).
The first scrubber began operation in December 1971. Total cost of the scrubber
system was $30 million, or $52 per kw.
The subbituminous coal burned at the Four Corners Plant, supplied by the
Navajo mine, has as its outstanding characteristic, an unusually high ash content
of nominally 22 pet. This results in a high dust loading in the flue gas
leaving the boiler, 12 gr/scfd. Initial operation of the plant using mechanical
collectors for fly ash removal resulted in dense plumes from the stacks, which
dispersed to restrict visibility over a large surrounding region. Since
scrubbers and ESP's have been operating, the plumes are greatly improved,
although still visible.
The Navajo coal averages 0.68 pet sulfur, which produces a sulfur dioxide
content of approximately 650 ppm. The fly ash alkalinity is low for a Western
coal, having only 5 pet CaO in the ash. However, the amount of calcium entering
the scrubber system is still relatively large, owing to the high percentage
of ash. The calcium is chemically equivalent to approximately 75 pet of the
coal sulfur content; and total alkalinity, including small amounts of Na20 and
MgO, exceeds 100 pet of sulfur equivalence.
A detailed description of the Four Corners scrubber installation is given
in the Appendix and in Figure 2. Flue gas from the air heaters enters the
venturi and then passes consecutively through a mist eliminator, a wet ID fan,
another mist eliminator, and a steam reheater. However, because of erosion
problems, the reheaters were removed and the units are operated as wet stacks.
There is no bypass. Turndown is 50 pet.
274
-------
Flue gas from
air heaters
To stack
Water spray
Mist
eliminators
Distribution
tank
Mist eliminators
Lime
Make up
water
Liquid transfer
tank
figure 2.
rubbers.
- Simplified flow diagram for the Four Corners fly ash
275
-------
Scrubber liquor is recycled back to the venturi and mist eliminator. The
system vas modified so that a bleed stream is taken from the recycle pump and
sent to the thickener. Before the modification, blowdown was drained by
gravity to the thickener. However, the gravity flow line was prone to plugging.
Lime is now added to the overflow launders of the thickener instead of the
centerwell, where some lime was lost by settling. The thickener and the fly
ash transfer tank are at the same elevation and a line between them is used for
water level control. The thickener underflow is pumped to a fly ash transfer
tank and then pumped to a settling pond. The sludge is reported to have good
settling properties. The pond is periodically dredged and disposed of in the
mine.
The key operating variables are a liquid-to-gas ratio (L/G) of 18 gal/1000
scf, a pressure drop across the scrubber tower of 18-20 inches of water, and a
pressure drop across the entire scrubbing system of about 28 inches of water.
The venturi recycle liquid has a pH of 3.2 to 3-5. The pH of the thickener
overflow liquid is maintained at about 7-5 by the addition of lime.
The operation of the venturi scrubbers has been satisfactory from the
standpoint of meeting the particulate removal goal of 99.2 pet. Past measurements
of sulfur dioxide removal, without supplemental lime, are reported to be in the
range of 30 to 35 pet. The present lime addition rate is equivalent to about
7 pet of the sulfur dioxide in the inlet flue gas. Therefore, it would be
expected to improve the sulfur dioxide removal by about 5 pet. The actual
sulfur dioxide removal efficiency has not yet been measured. However, a long-
range testing program with the primary purpose of reducing sulfur dioxide
emissions, will be implemented by Arizona Public Service Company, since the
existing level of sulfur dioxide emissions will not meet new sulfur dioxide
limits. A secondary purpose of the test program will be to reduce scaling.
The testing program to be implemented will modify one scrubber system to
test several variables. The variables to be tested are various liquid-to-gas
ratios and suspended solids. Additional spray nozzles will be installed in
the sorubber and an additional rubber-lined recycle pump will be added. The
sulfur dioxide removal efficiency will be investigated as a function of L/G.
The level of suspended solids will be maintained at 6 pet to provide seed
crystals for precipitation of gypsum. Lime will be added for pH control.
Scrubber operating costs are not separated from the plant operating cost,
and hence, are not available. The major operating requirements are electrical
power equal to 3 to k pet of generating capacity, an estimated water usage
of 5-9 acre-ft/MW/yr, manpower including 8 operators plus maintenance and
supervision, and Ik tons of lime per day for control of pH.
Availability for the scrubber system is currently estimated at about 80 pet. •
This is the fraction of the time that a boiler is operating or could be operating
that the scrubber modules are also operative. Since there is no bypass, this
level of availability involves appreciable loss in power generation. With two
scrubbers per boiler, lost generation can involve either reduction in load when
one scrubber is inoperative or complete boiler shutdown when both scrubbers are
inoperative.
276
-------
The operating problems have "been scale and solids buildup in the scrubber
would break off and clog the gravity-feed blovdown line. A severe erosion
problem in the venturi opposite the plumb bell vas resolved by installing silica
carbide bricks. The bricks are reported to be very effective in resisting
erosion. A problem of erosion on the tangential nozzles resulted in by-passing
dirty flue gas to the wet fan. This, in turn, resulted in severe fan erosion.
The problem was resolved by adding a stainless steel wear plate to the tangential
nozzles. Additional problems are corrosion and leakage in the recirculatory
lines and vessels and reheaters, vhere coatings on carbon steel has failed, on
the mist eliminators and on the ducting leading into the stack. The blowdovn
line plugging has been resolved by using a forced bleed from the recycle pump.
A nanually-operated valve has been inserted in the old existing gravity blovdown
line for emergencies. The reheater corrosion problem was resolved by removing
the reheaters and using a wet stack. However, the stack liner and mortor were
not acid resistant and, consequently, had to be replaced with the proper liner.
Scaling has occurred on most wetted surfaces, and it is not yet under
control. Control measures include the use of an appreciable amount of blowdown
(open-loop operation), an increase in the percentage of fly ash solids in the
reeirculating scrubbing liquor (from 2 to 6 pet), and addition of a new lime add
System to maintain the pH at 7-5 in the thickener. It is not clear why this pH
aijustment insures improved control of scaling, since in a system where the
state of oxidation is high, with most dissolved sulfur present as sulfate,
calcium sulfate would not be expected to be precipitated in the thickener by the
rise in pH?.
The measures being investigated for scale control at Four Corners, if
successful, should find wide application in scrubbing applications involving
lew-sulfur Western coals.
SftSD FORKS ENERGY RESEARCH CENTER
The Energy Research and Development Administration's Grand Forks Energy
tesearch Center (GFERC) has investigated the fly ash alkali sulfur dioxide
scrubbing system since 1971. Testing has been performed on a 130 scfm pilot
scrubber. The principal objectives have been to determine sulfur dioxide
KBOTal efficiencies and calcium sulfate scaling rates as a function of sulfur
dioxide level, fly ash add rate, alkali in the fly ash, supplementary lime
requirements, level of recirculated suspended solids, liquid to gas ratio,
SBunt of makeup water and total dissolved solids. Past results have been
l&lished in three previous papers"'^'10. Research is continuing under funding
from the Environmental Protection Agency, investigating the effect of high
Jsrels of total dissolved solids on sulfur dioxide removal efficiencies and
tflcium sulfate scaling.
The present GFERC scrubber (Figure 3) is a 130 scfm flooded disk venturi
by an absorption tower containing conical "rain and drain" trays.
&6Bsure drop across the scrubber can be controlled by adjusting the height of
tie flood disk. The conical trays were installed as a modification to increase
-------
S(>2 injection
Gas furnace
NJ
^1
00
Water cooled
heat exchanger
Floating weir
Rain and
drain tower
Drip leg
: rj
To stack
,
b
ID fan
Fly ash
^K
.>
r
o
]
o
r* —
^ L
o
*v. °
r ;
o
0 .,
Mix tanks
Fiaure 3. - Pilot olant scrubber. Grand Forks Enerav Research Center.
-------
removal levels so that the sulfur dioxide removal observed would Toe a function
of the solution characteristics only and not of scrubber design. Installation
of the conical "rain and drain" trays increased the removal efficiency by 5 pet,
from 83.3 to 88.1 pet under identical operating conditions.
The GFERC scrubber system is "closed loop." Recirculating scrubber liquor
is lost from the system only as mist, which is equivalent to about 0.8 acre-
ft/MW/yr, or as liquor in sludge. Efficient mist elimination is accomplished by
passing gas through both a cyclone and a stainless steel wire mesh. Water lost
by evaporation is replaced, but mist loss is not replaced during the course of a
one-week test. The scrubber liquor is returned to a series of two fly ash mix
tanks equipped with overflow weirs. The overflow from the second mix tank flows
to a settling tank where calcium sulfate precipitates and unreacted fly ash ±s
allowed to settle. A floating overflow weir in the settling tank provides the
scrubbing liquid to the flooded disk.
Early experiments at GFERC indicated that a large increase in total dissolved
solids, primarily sodium and magnesium sulfates, during the approach to steady
state operating conditions significantly increased the scrubbing efficiency.
Since the fly ash derived from some Western coals are known to contain significant
amounts of soluble sodium and magnesium, it is probable that high concentrations
of these species will result after long-term operation of a full-scale scrubber
employing the fly ash alkali scrubbing process. Current experiments at GFERC
are designed to investigate the properties of scrubber solutions that are high
in sodium (0.5 to 10 pet) and magnesium (0.5 to 10 pet). The objectives of the
current tests are to determine sulfur dioxide removal and scaling rate using a
solution concentrated in sodium and magnesium and low in suspended solids (high
levels of suspended solids, 6 to 12 pet, are common practice for scale control
in many western scrubbers). The fly ash used in these tests contained high
sodium and magnesium and was produced by pc-firing of Beulah, North Dakota
lignite. Scrubber operating conditions kept constant for all test runs were:
inlet sulfur dioxide level of about 8UO ppm (a typical Western lignite containing
about 0.8 pet sulfur), inlet flue gas temperature of 350° F, liquid temperature
of about 120° F, absorber tower pressure drop of about 13 inches of water.
Scaling rates reported represent the rate of weight increase in grams per
hour observed in a 3 ft k inch length of 1/2 inch I.D. pipe in the return line
from the scrubber to the mix tanks. The test position chosen was a point of
maximum scaling, and trends in the values observed were found to be well correlated
with operating variables.
Tests were performed at liquid to gas ratios of 23, U5, and 75 gal/1000
scf. The CaO/SC>2 stoichiometric ratio was maintained at 1.2, sodium concentration
at about 3.0 pet, magnesium concentration at 1 to 2 pet. The pH of the liquor
pumped to the absorber tower varied from 5-0 to 5-5. Previous experiments
indicated only a marginal effect when L/G was increased. However, under the
conditions of high sodium and magnesium, removal efficiencies were affected very
significantly. The removal efficiencies and fly ash alkali utilizations are
tabulated in Table 3.
279
-------
TABLE 3 SULFUR DIOXIDE REMOVAL EFFICIENCIES AND FLY ASH
ALKALI UTILIZATION AS A FUNCTION OF L/G. CaO/SOg = 1:2
L/G
23
H5
75
Removal Efficiency (pet)
81.0
88.2
98.1
Alkali Utilization (pet)
66
72
80
The stoichiometric ratios of calcium oxide to sulfur dioxide investigated
were 0.6, 1.2 and 2.0. These ratios correspond to particulate loadings of 2.0
gr/scf, U.O gr/scf and 6.7 gr/scf. Operating conditions were as described
above, with an L/G of ^5. The sulfur dioxide removal efficiency, fly ash alkali
utilization, and scaling rate are tabulated in Table U. Removal efficiencies
are shown in Figure 6.
TABLE k SULFUR DIOXIDE REMOVAL EFFICIENCIES, FLY ASH
ALKALI UTILIZATION AND SCALING RATE AS A
FUNCTION OF STOICHIOMETRIC RATIO, CaO/S02
CaO/S02
0.6
1.2
2.0
Removal
Efficiency (pet)
63.3
88.2
98.0
Alkali
Utilization (pet)
100
72
U9.1
Scaling
Rate (gm/hr)
1.68
2.7
3.0
In the above tests, some difficulty was experienced in removing the suspended
solids to produce a "clear" liquid, even with the addition of sodium aluminate
to the scrubber solution as a coagulant. The scaling rate of 1.68 gm/hr was
observed at a suspended solids concentration of 0.13 pet, 2.7 gm/hr at 0.18 pet,
and 3.0 gm/hr at 0.2U pet. Previous experience has shown that 10 gm/hr is a
high scaling rate, and 0.2 gm/hr is a low scaling rate.
After the foregoing tests, the absorber tower was again modified, this time
for the purpose of providing greater control of the pressure drop under conditions
of severe scaling. The change involved attaching one set of cones, those directing
flow from the center outward, to the standpipe of the flooded disk. Thereafter,
movement of the standpipe varied the spacing between the convex and concave
conical trays as well as the spacing of the flooded disk venturi. Thus, as
scale buildup occurred on opposed surfaces, all such surfaces could be moved
further apart to maintain a constant pressure drop. A further effect of the
change was to distribute the pressure drop more evenly throughout the scrubber
tower. This last effect was believed to be responsible for a further increase
observed in scrubber efficiency, from 88.2 to 92.9 pet, which probably occurred
because of a more efficient use of energy in redispersing droplets of scrubber
liquor throughout the tower. All of the removal data given below are offset
from former data due to this increased efficiency.
280
-------
Scaling rates and removal efficiencies were next investigated as a function
of sodium concentration. The sodium levels investigated were 0.17 pet, 0.66
pet, k pet, and 9-3 pet. The 0.17 pet represents sodium leached from the fly
ash during a 3-day test period; no make-up sodium was added. The magnesium
concentration was kept constant at 1 to 2 pet, L/G was ^5, calcium oxide to
sulfur dioxide stoichiometric ratio was 1.2, and other operating conditions were
as described previously. The results are tabulated in Table 5- A typical
solution analysis at each sodium level is listed in Table 6.
TABLE 5 SULFUB DIOXIDE REMOVAL AND SCALING RATE AS A
FUNCTION OF SODIUM CONCENTRATION
Sodium Concentration (pet)
0.17
0.66
h.Q
9.3
S02 Removal
Efficiency (pet)
95-2
92.9
93
96
Scaling
Rate (gm/hr)
5-8
3.51
2.7
.0
Suspended
Solids (pet)
0.066
0.07k
0.17
0.83
TABLE 6 TYPICAL SOLUTION ANALYSIS AT SODIUM LEVELS OF
0.17 PCT, 0.66 POT, h.O PCT, AND 9-3 PCT
Species:
Ca (ppm)
Mg (pet)
SOl^ (pet)
Percent Sodium: 0.17
702
1.30
7-06
0.66
631
1.33
7-56
h.O
6H3
1.3
15-0
9-3
762
1.7
26.0
The solids settling properties were observed to degrade as the ionic strength
of the scrubber solution increased. This phenomenon has been observed previously
in EPA laboratory testing on dilute double alkali systems, and by Arthur D.
Little, Inc.11 in laboratory and pilot plant work on dilute and concentrated
double alkali systems. Factors reported to influence the solids settling properties
are reactor configuration, concentration of soluble magnesium and iron, and the
concentration of soluble sulfate. In this investigation, the concentration of
magnesium and iron were relatively constant. However, the level of soluble
sulfate varied along with sodium level, due to the addition of sodium as sodium
sulfate. The solids settling characteristics degraded correspondingly. A
vacuum filter is being installed on the GFERC scrubber so that the level of
suspended solids may be controlled more precisely.
It can be seen from Table 5 that the rate of scaling decreased as the
sodium concentration increased. The absence of scale formations at the 9-3 pet
sodium level is thought to be a function of sodium and not due to the higher
(0.83 pet) level of suspended solids, since past work at similar levels of
suspended solids (low sodium) resulted in scaling rates of 1 to 2 gm/hr. The
stack flue gas was also tested to determine if sulfate was being lost in the
281
-------
mist. If sulfate was lost in the mist at a rate equal to or greater than that
being absorbed into the scrubber solution (assuming constant liquid volume in
the system), scaling would not "be expected to occur. Extensive testing indicated
this did not occur, and thus, the absence of scaling is concluded to be a result
of the high sodium concentration.
It can also be seen from Table 5, that an increase in the level of total
dissolved solids did not have a significant effect on sulfur dioxide removal,
which contradicts previous results. The current result showing no effect on
removal was observed after the modification of the scrubber by installation of
"rain and drain" trays, which greatly increased gas-liquid contact in the scrubber.
The conclusion to be drawn is that a high ionic strength in terms of sodium and
magnesium sulfates increases removal for a scrubber configuration providing
minimum contact-residence time (the flood disk venturi alone) but that it does
not increase removal for a scrubber providing a maximum of contact-residence
time (the venturi plus trays). The results would further indicate that the
scrubber solution having low ionic strength had a sufficient equilibrium capacity
to absorb essentially all of the entering sulfur dioxide (at 8^0 ppm and
L/G = 1<5), but that this capacity was not fully utilized without the increased
residence-contact time. On the other hand, the scrubber liquor having high ionic
strength was indicated to have a greater affinity for rapidly absorbing sulfur
dioxide such that essentially all entering sulfur dioxide could be removed with
a short residence-contact time. Thus, the final conclusion is that sulfur dioxide
removal in ash alkali scrubbing can be materially improved by either an increase
in ionic strength or an increase in residence-contact time, but that a substantial
increase in either can mask the effect of the other.
State of oxidation, of sulfite to sulfate, was generally high for all test
conditions (98 to 99 pet sulfate). The percentage of sulfite was, however,
higher in the test run at 9.31* pet sodium (2 pet sulfite) than in any other
test.
Future tests will investigate the solution effects of varying the magnesium
concentration at constant sodium levels and of scrubbing at low pH (below pH 3).
In the present investigation, the level of magnesium was maintained at a constant
value.
The GFERC is also working under a cooperative agreement with the Square
Butte Electric Cooperative (SBEC), Minnesota Power and Light Company (MP&L), and
Combustion Equipment Associates (CEA), to investigate the fly ash alkali scrubbing
process using a 5*000 acfm (saturated) pilot scrubber. The objectives of the
cooperative agreement are listed in the SBEC section.
Work is also planned on properties of scrubber waste from ash alkali scrubbing.
The ideal properties for a scrubber waste are:
- absence of toxicity
- low soluble solids content
- low moisture content
- non-thixotropicity
- high compressive or bearing strength
282
-------
Solid waste studied will be generated from operation of the 130-sfm scrubber and
from SBEC's 5»000 acfm (saturated) pilot plant. Solubilities will be determined
for the major soluble elements and for selected trace elements under conditions
representative of disposal.
1 *?
Minnesota Power & Light - Clay Boswell (Cohasset) and Aurora Stations
Minnesota Power and Light operates Krebs-Elbair spray impingement scrubbers
for particulate control on two 58 MW boilers at the Aurora Station and on one 350
MW boiler at Cohasset.
The Krebs-Elbair scrubber consists of a stainless steel box containing
nozzles that direct a high pressure spray against baffles. The baffles consist of
either vertical rods or a punch plate. The liquid is atomized when the spray
impinges upon the baffles, and the resulting turbulance promotes an effective
scrubbing of particulates. The Krebs-Elbair scrubber is designed to substitute
power input from the high pressure spray for flue gas pressure differential, and
it operates with only a k-incti pressure drop across the scrubber system. This
design supposedly affects a net operational savings in power and cost when
compared to scrubbers employing relatively large pressure drops in the gas
stream. An additional important feature is a nozzle tree design which permits
sections of the nozzles to be removed for maintenance without scrubber shutdown.
An apparent design limitation of MP&L's scrubbers is the erosion and plugging of
the high pressure nozzles at other than very low levels of suspended solids. As
previously stated, the scrubbers are designed primarily for particulate removal,
although sulfur dioxide is removed because of the alkaline character of the fly
ash particulate.
The coal burned at both the Aurora and Cohasset stations is a subbituminous
from the Big Sky Mine, located near Colstrip, Montana. The coal typically contains
1.0 pet sulfur, 10 pet ash, and 9 to 13 pet CaO in the fly ash. The scrubber
inlet particulate loadings are 2 gr/scfd at the Aurora station and 3 gr/scfd at
the Cohasset station. A typical scrubber inlet sulfur dioxide level is about 800
ppm.
A detailed description of the MP&L scrubbers is given in the Appendix. A
process diagram is shown in Figure h. Flue gas from the air heaters passes
through three concurrent sprays: a quench spray, high pressure spray, and the
post humidification spray. The mist eliminator is between the high pressure
spray and the post humidification spray. The post humidification spray washes the
ID fan, which discharges flue gas to a wet stack. Scrubber liquor is pumped from
a seal tank at the bottom of the spray chamber to two clarifiers. Overflow from
the clarifiers is combined with makeup water and pumped back as spray. The ID
fan is washed only with makeup water. Slowdown from the clarifiers is pumped to
an eighty-acre ash pond. The scrubber operation is not closed loop. The Aurora
station does not have clarifiers, and instead uses the clear liquor from the ash
pond. In all other respects, the Aurora station is similar to that at Cohasset.
283
-------
Make up water
To stack
£>X Post humidification
" ^~ spray
Mist eliminator
Punch plate
Bottom ash pond
Figure 4. - Simplified flow diagram for the Clay Boswell station
particulate scrubber.
284
-------
The key operating variables at "both stations are L/G of 8.3 gal/acf and a
total gas stream pressure drop of about It inches of vater. The high pressure
spray enters the scrubber at 200 psi. The pH of the liquid entering the scrubber
is typically it.It and about it.O when leaving the scrubber. Particulate removal
efficiency is about 98 pet at Aurora and about 97 pet at Cohasset. The sulfur
dioxide removal is about 20 pet at Aurora and about 15 pet at Cohasset.
Operating costs are not available. Operating requirements are electrical
power equal to 0.86 pet of generating capacity, water requirements of about 6.5
acre-ft/MW/yr at Cohasset and 30 acre-ft/MW/yr at Aurora, and a high labor
requirement for maintenance and operation.
The scrubber availability was not obtainable in terms of effective scrubber
operation during times that boilers were operational. However, little down-time
on boilers would be expected using this scrubber process, since many scrubber
problems can be repaired without boiler shutdown. Massive scaling, plugging, or
problems involving the wet ID fan would, however, necessitate boiler shutdown.
The major problems that have occurred are stack gas mist carryover, scaling
in the scrubber and liquid circuit, and heavy scale deposits on the wet ID fan
and mist eliminators. The scaling problem was more severe at Cohasset, which
operates close to its rated load and is restricted on the amount of blowdown.
Eovever, a seasonal variance had been obtained which allows greater blowdown,
and the scaling and plugging problems have been less severe. About U20 gpm of
clarifier underflow is pumped to the 80-acre fly ash pond. About 1000 gpm of
sulfate-laden overflow from the clarifier is drained to the coal pile sump where
it is diluted with drainage from selected plant drains, demineralizer waste
rater, boiler blowdown, stack drainage, and drainage from the coal pile and then
pumped to the bottom ash pile pond and diluted with bottom ash slurry water.
Some sulfate removal by precipitation is experienced. Overflow from the ash pond
is pumped to the cooling water intake for units 1 and 2 and subsequently discharged
into the Mississipi River.
The problem of stack gas mist carryover has been shown to be solvable by
the use of a low velocity stack. A low velocity stack was simulated by breaching
the brick liner of an existing high velocity stack. The flue gas could then
flow through the annulus of the stack in addition to the brick liner and thus,
in effect, simulate a low velocity stack. Subsequent tests on the modified
stack indicated that mist carryover would not be a problem. An unexpected
result of the low velocity stack tests was that the overall sulfur dioxide
removal was increased from about 15 pet to 37 pet.
The increased sulfur dioxide removal in the wet stack was attributed to the
TOt ID fan and the modified low velocity stack. The water used to wash the ID
fan is shattered into droplets which, depending on their size, would rapidly
fall out or are transported to the stack. Depending on the transport velocity of
the droplets, they could be carried out of the stack (which they were not), fall
to the bottom of the stack, or be absorbed onto the walls, with some drops
remaining in a dynamic suspension until becoming large enough to fall to the
285
-------
"bottom of the stack. A situation has thus been created in which a water aerosal
is continuously and intimately mixed with flue gas and fly ash aerosal. As a
consequence, an additional IT to 22 pet sulfur dioxide removal efficiency was
obtained in the wet stack and the overall removal efficiency was increased to
about 37 pet. The liquid that was drained from the stack to the bottom ash pond
had a pH of about 3.
Minnesota Power and Light had planned on two TO MW retrofit scrubber
installations on units 1 and 2. However, MP&L now tentatively plans to use
ESP's on the two TO MW boilers and the Krebs-ELbair scrubber on the 350 MW unit
and use a common low velocity stack. With all three boilers operating, the flue
gas from the two TO MW boilers would provide reheat to the 350 MW scrubber exit
flue gas.
Montana Power Company Pilot Plant
A pilot plant program investigating ash alkali scrubbing was undertaken in
19T3 for Montana Power by Bechtel Corporation in cooperation with Combustion
Equipment Associates who designed the scrubber. Arthur D. Little, Inc. also
participated in the test program, and sulfur dioxide and particulate testing
services were provided by York Research Corp. Based on generally favorable
results in the pilot plant program, Montana Power is employing ash alkali scrubbing
in a new 362 MW pc-fired generating unit at Colstrip which started operation in
September 19T5. A similar unit is scheduled for startup in 19T6 or 19TT-
The information presented on these tests was obtained primarily from Dr.
Carlton Grimm of Montana Power Company, and from a report published by Combustion
Equipment Associates-^. The present description of the tests as pieced together
from the various sources is solely the responsibility of the writers of this
paper.
The 3,000 acfm pilot plant (Figure 5) consisted of a venturi section for
particulate removal, a spray tower for sulfur dioxide removal, followed by mist
eliminators and a reheat section. Provisions were made for alkali makeup as
or lime and the use of cooling tower blowdown as makeup water.
The principal objectives of the study were to:
1. Demonstrate CEA's guarantees on sulfur dioxide and
particulate removal.
2. Determine level of supplemental alkali required, if
any, for sulfur dioxide removal.
3. Optimize variables influencing system performance.
1». Investigate use of cooling tower blowdown as a
demister wash spray.
286
-------
Absorber
N)
OO
To stack
t
Steam
reheater
\' V \
Recycle
tank
•Q-
Fan
Figure 5.
Sludge to ash pond
Pilot plant scrubber, Montana Power Company
Make up water
formulation
_L
Alkali make up
tanks
-------
The goal for particulate removal was an outlet dust loading of 0.03 gr/scf.
Results at an inlet dust loading of 2 gr/scf indicated that the 0.03 gr/scf exit
loading could be met at a venturi pressure drops of approximately 12 inches of
H20, and 0.02 gr/scf at about 17 inches of H20. A pressure drop of 17-inch H20
was selected for the h to 6 gr/scf anticipated in the full scale units.
The sulfur dioxide removal requirement was to comply with a standard of 1 Ib
S02/MM Btu, or a level of U25 ppm. Compliance required removals in the range of
50 to 60 pet. This was accomplished with fly ash alone (without supplemental
alkali) supplied at a dust loading of 2 gr/scf. At this dust loading, however,
a pH of U in the scrubbing liquor entering the venturi was judged lower than
desirable for scale control. Using a simulated grain loading of k gr/scf, a
higher pH of 5 to 6 was achieved and the removal criterion was also met. An
approximate plot of sulfur dioxide removal efficiency as a function of stoichiometric
ratio alkali/S02 (inlet) is shown in Figure 6, for ash alone and with lime
added.
A suspended solids level of 12 pet was thought to be adequate for sulfur
dioxide removal and scale control, but higher levels would furnish more residence
time for the ash to react and provide additional nuclei for precipitation.
Suspended solids level was found to have an important positive effect on
sulfur dioxide removal in the range of 3 to 12 pet. Increasing L/G in the spray
tower also increased sulfur dioxide removal. L/G and pressure drop in the
venturi had little effect on sulfur dioxide removal.
Tests were run in which NagCC^ was added for pH control. Sulfur dioxide
removal was improved by 60 pet or more of the stoichiometric equivalent of the
NaoCOo added. There are no plans to use soda ash in the full scale units at
Colstrip.
In other tests, lime was added to supplement the ash alkali. Sulfur
dioxide removal was improved by 60 pet or more of the stoichiometric equivalent
of lime added. (See Figure 6 for approximate results.) Use of lime is planned
in full scale units to control pH during periods when fly ash loadings are less
than k gr/scf or when fly ash reactivity decreases.
Scaling in the scrubber and liquid circuit was controlled adequately,
mainly by the recirculation of ash. Some fouling did occur in the wet-dry zone,
and loose scale deposit accumulated in the reheater.
Major emphasis was given to testing the chevron mist eliminators to establish
their effectiveness and the wash conditions required to prevent plugging. Three
approaches indicated that there was little measurable liquid entrainment through
the dsmister. First, direct measurement of the degree of saturation of the flue
gas indicated no carryover. Second, outlet particulate loadings with variable
suspended and dissolved solids in the liquid showed no increase in outlet loading.
Third, outlet particulate loadings did not change upon increasing or reducing
demister wash sprays. However, deposits did form on the reheater, indicating
that some carryover was occurring.
288
-------
K)
OO
too
80
o
k.
Q>
Q.
60
Ul
a:
ftl
O
V)
40
20
0.5
Montana Power (fly ash and lime)
Montana Power (fly ash only)
NSP (fly ash and limestone)
GFERC (fly ash only )
1.0
2.O
3.0
4.0
STOICHIOMETRIC RATIO —— INTO SCRUBBER
SO?
Figure 6. - Sulfur dioxide removals In pilot plant tes'ts on alkaline fly ashes.
-------
Simulated cooling tower blowdown was found to "be unacceptable for washing
the mist eliminators, because of scaling that occurred. To combat this scaling,
a Koch bubbler tray was added ahead of the mist eliminator and reheater. Fresh
makeup water used to wash the mist eliminator dropped into this tray, and re-
entrainment of this relatively clean water from the Koch tray lowered suspended
and dissolved solids in the mist by dilution and "exchange," aiding significantly
in keeping the mist eliminator clean. Plans call for using this feature in the
full scale units.
The level of dissolved solids in recycled scrub liquor was influenced by
the amount and solubility of cations derived from the ash and by the amount and
quality of makeup water added. Analysis at this end of a prolonged period of
closed loop operations indicated that dissolved solids reached 26,000 ppm.
Dissolved ion concentrations were as follows:
Magnesium U.OOO
Chlorine ^00
Sulfate 18,000
Sulfur trioxide 3,000
Calcium ^00
The principal findings of the Montana tests were that acceptable sulfur
dioxide removal could be achieved using fly ash alone as the source of alkali ,
and acceptable particulate removal could be obtained. Supplementary alkali
could be used to control pH if fly ash quantity or quality were reduced. The
optimum L/G for spray tower operation was determined to be in the range of 15 to
20 gal/1,000 scf.
Scale control .was achieved principally through the recirculation of ash
solids at 12 pet and above. Control of pH also aided in scale control. Other
measures found helpful were fresh water washing of the mist eliminator and the
use of the Koch bubbler tray. Cooling tower blowdown was found unacceptable as
a source of makeup water for washing this mist eliminator. Operation of the
full scale scrubbing unit at Colstrip has been considered successful to date.
lit
Northern States Power Company - Black Dog Pilot Plant
A study on fly ash alkali scrubbing was undertaken by Northern States
Power, at their Black Dog Plant near Minneapolis, to establish the design of
scrubbers for the Sherburne County Generating Plant where two 680 MW units are
scheduled for completion in 1976 and 1977. Other parties involved in the test
program included Combustion Engineering, the scrubber vendor, and Black and
Veatch, consulting engineers.
The test facility was a marble bed scrubber with demister and reheat
capability, having a capacity of 12,000 acfm. Attendant equipment included a
reaction tank, thickener and ash pond, makeup and limestone tanks. Arrangement
of equipment is shown in Figure 7-
290
-------
Flue gas
Limestone
slurry make up
Primary
contactor
Air
Figure 7. - Pilot plant scrubber, Northern States Power.
-------
The principal objectives of the test program were to:
1. Demonstrate capability for an S02 removal of 50 pet
or greater operating on alkaline ash only or on
ash and limestone.
2. Demonstrate the guaranteed particulate removal capability
to O.OH grams/scfd.
3. Demonstrate reliable operation with acceptable
maintenance.
h. Determine optimum operating conditions, including
blowdown requirements, and instrumentation and
control functions.
5. Demonstrate reliable operation without deposit for
mation on the components.
6. Demonstrate acceptable continuous operation over
a 30 day period.
To obtain the guaranteed dust removal of 0.0^ gr/scfd outlet dust loading,
or 99 pet removal, it was necessary to modify the marble bed scrubber by installing
a venturi rod section (commonly called the primary contactor) at the inlet. The
venturi rod section increased the total pressure drop from 6 to over 12 inches
of water and the particulate criterion was met. Optimum system L/G was about 30
gal/1,000 acf. A demonstration test of the two-stage scrubber system for 50
continuous days was completed with 99 pet availability.
A series of tests were made wherein inlet sulfur dioxide and supplementary
limestone were varied while the input fly ash remained relatively constant. The
sulfur dioxide was varied from U82 ppm to 919 ppm and the limestone from 9 pet
to 100 pet of stoichiometric. The sulfur dioxide removal efficiency varied from
k8 to 88 pet under these conditions. The relationship between the stoichiometric
ratio of input alkali to sulfur dioxide and the removal efficiency is shown in
Figure 6. The sulfur dioxide removal criterion was met with as little as 5 pet
stoichiometric limestone added. The calcium in the fly ash played a major part
in the removal of sulfur dioxide, representing 70 to 80 pet of the alkali reagent.
Some scaling and plugging which occurred in the pilot program was remedied
in the course of testing prior to the endurance test. Heavy mud deposits up to
two inches thick appeared on the mist eliminator and scrubber wall surfaces,
requiring shutdown every two days for washing. Deposits in both the mist
eliminator and reheater were similar in chemical composition to the spray water
solids, consisting primarily of fly ash with CaS03, CaSO^, and CaCOg. Deposits
on the ID fan were chemically similar but physically more amorphous. Hard
calcium sulfate scaling occurred on overflow pots. A redesign of the mist
eliminator washing equipment was made to prevent plugging at that point. In
addition, three modes of sulfate scale control were utilized. Ash solids were
recirculated to provide ash and gypsum seed crystals, with a 10 pet level of ash
292
-------
found to be optimum. Supplementary limestone was reduced to a level of giving
adequate sulfur dioxide removal while minimizing scale, with a level of approximately
15 pet of stoichiometric or less found to be optimum. A level of 30 pet
stoichiometric was used in the endurance test. Oxidation of sulfite to sulfate
was increased by bubbling air into the reaction tank to help prevent supersaturation
with calcium sulfite in the scrubber proper. Oxidation level was increased from
36 pet to 98 pet and above with four times stoichiometric air. Good experience
was reported with rubber-lined process equipment and fiberglass spray headers.
slurry nozzles did not show wear or deterioration after 1700 hours of operation.
After kO days of operation, the following concentrations of dissolved
species in the spray water were measured:
PPM
Sulfite 0
Sulfate 30,000
Chloride 550
Nitrate kOO
Calcium UOO-500
Magnesium T»000
Silica 200
Sodium 200
Pacific Power and Light - Dave Johnston Plant '
The Dave Johnston Plant at Glenrock, Wyoming has one 330 MW pc-fired boiler
equipped with three parallel Chemico venturi scrubbers. The initial capital
investment was $8 million, or $2U/kw. A reported $5 million has beea spent on
improvements, bringing total cost to $39/kw. Startup was in April 1972.
Coal burned at the Dave Johnston Plant is Wyoming subbituminous coal from a
captive mine. Sulfur content is 0.5 pet, resulting in a sulfur dioxide content
of 500 ppm in the untreated flue gas. Coal ash content is 12 pet, and the
calcium oxide content of the ash is approximately 20 pet. The calcium is
chemically equivalent to 275 pet of the sulfur content. Inlet dust loading is k
gr/scfd.
A detailed description of the Dave Johnston scrubber installation is given
in the Appendix and Figure 8. Flue gas from the air heaters enters the venturi
and then passes sequentially through mist eliminators, to a wet ID fan, and on
to a wet stack. Ho reheat is used; there is no bypass. Turndown is to approximatel:
30 pet of scrubber design capacity.
Scrubbing liquor is continuously recycled from the bottom of the venturi
scrubber back to the plumb bob and to the deflector surrounding the bob that was
installed to prevent solids buildup. Slowdown from this loop is pumped directly
to two fly ash settling ponds; no thickener is used. Overflow from the settling
pond is sent to a clear pond and then pumped back to the recycle loop. Ash is
dredged from the settling ponds once each year and is hauled away for landfill.
293
-------
Flue gas from
air heaters
Mist eliminators
pond ^m/
ash
Figure 8. - Simplified flow diagram for the Dave Johnston fly ash scrubbers.
-------
Key operating variables are an L/G of 13 gal/1,000 acf and a total pressure
drop of 15 inches of water. The pH leaving the scrubber is 5, without supplemental
lime.
Operating costs are not available. Major operating requirements are
electrical power equal to 2.3 pet of generating capacity, an estimated water
requirement of 3.6 acre-ft/Mtf/yr, lime for control of pH, and manpower.
Particulate removal efficiency is over 99 pet, meeting the design goal of
O.OU gr/scf. Preliminary values for sulfur dioxide removal are ho pet without
lime and somewhat higher with lime addition.
Availability is characterized by PP&L as being less than adequate for
utility use, but no company-sanctioned percentages are available. Availability
depends on the amount of blowdown and fresh water irrigation that are employed.
Operation is characterized as "intermittent open loop," meaning that
operation with a minimum of blowdown is attempted as the normal mode of operation,
with much larger amounts of blowdown and makeup used periodically to irrigate
the system. Cooling water blowdown is mixed with service water (North Platte
River water) and used for washing the ID wet fan.
Past major operational problem was scaling. At present, it is reported
that scaling of the scrubber is still a major problem and an occasional fresh
water irrigation is required. However, the previous detremental effects of wet-
dry interface scaling have been minimized by the application of lignosulfonate,
and is now not considered a major operational problem.
Future test plans are concerned with minimizing scale formation. The
scrubber will be operated continuous with the recycle liquor pH maintained at
5.5 to 6.0, and scaling rates and cleaning requirements will be determined. The
pH will be controlled using hydrated lime and feed rates are anticipated to be
about 1,000 Ib/hr. If the test is successful, a major development program will
be implemented to design and install a permanent lime system utilizing pebble
lime to maintain the pH at a constant level. The hydrated lime and future
application of pebble lime will be to the scrubber vessel.
16
Public Service of Colorado - Valmont, Cherokee, and Arapahoe Stations
The Public Service of Colorado has five turbulent contact absorbers (TCA)
scrubbers containing thirteen modules installed on pc-fired boilers. The modules
were designed by Universal Oil Products primarily for particulate removal.
However, because of the alkali present in the fly ash, some sulfur dioxide is
also removed. The modules consist of three stages of mobile packing, commonly
called "ping pong balls," with the spray directed downward through the mobile
packing, countercurrent to the gas flow. A booster fan directs the flue gas to
the scrubber where it is presaturated and then passes through the mobile "bed,^
then to chevron mist eliminators and on to a reheater. Reheat on most units is
accomplished by heating the flue gas with steam coils; the Cherokee No. h unit
uses externally heated air. All units have by-passes. Typical turndown capability
is from Uj to 105 pet of the rated scrubber capacity. Detailed descriptions of
the installations are given in the Appendix and Figure 9-
295
-------
Flue gas to reheater
Flue gas from
electrostatic
precipitator
I
Surge
tank
Clear effluent
discharge
*i*/*v fls.» **.•?*.+ ***'
'•»'*'*•*»«# * '•- *-• *j
•///;.'.•.:•/;;:-;•/::•
Make up
water
Mist eliminator
Sludge pond
Ping
pong
balls
Figure 9. - Simplified particulate scrubbers at Valmont, Cherokee
and Arapahoe Stations, Public Service of Colorado.
296
-------
Public Service of Colorado's installation of the TCA scrubbers represents
the last stage of particulate removal in a system which also uses mechanical
collectors and electrostatic precipitators. All of the scrubber-equipped boilers
are still serviced by the previously installed mechanical collectors and ISP's
At the Valmont Station, flue gas from the mechanical collector on a 196 W
boiler is split into two equal streams with one stream sent to a scrubber and
the other stream sent to an ESP. The Cherokee Station, which has four boilers
with a gross capacity of 115 MW, 170 Mtf, and 375 *W, has a particulate removal
scheme in which the flue gas passes through, in series, mechanical collectors,
ESP's, and then through scrubbers. The Arapahoe Station has a similar scheme.
The coal burned at the Valmont and Arapahoe Stations is Wyoming subbituminous
and has a calcium content of 0.6 pet and an ash content of 5.2 pet. The calcium
oxide content of the fly ash is about 20 pet, which is chemically equivalent to
about 99 pet of the sulfur content. The Cherokee Station bums Colorado
subbituminous coal which has a sulfur content of 0.7 pet and an ash content of
9.U pet. The calcium content of the fly ash is about 5 pet, which is chemically
equivalent to about 38 pet of the sulfur content. The dust loading at all the
units range from 0.1+ to 0.8 gr/scfd. The sulfur dioxide concentration at all
the units is nominally 500 ppm.
Key operating variables at the Arapahoe and Cherokee Stations are a L/G
of 5^-58, a total pressure drop of 10 to 15 inches of water, an average pH of
7.0 entering the scrubbers and a pH of 2.8 to 3 leaving the scrubber. The
scrubber liquor is recycled from the bottom of the scrubber to the spray header
above the mobile bed. At the Arapahoe Station, slurry blowdown is pumped to an
ash settling pond after adjusting the pH of the liquor to 6.0 to 9.0 with lime.
The settled fly ash in the pond is dredged periodically and used for landfill.
Clear effluent from the ponds is discharged under permit from the State of
Colorado. At the Cherokee Station, slurry blowdown is neutralized and clarified
and then mixed with the ash pond overflow for discharge. Scrubber operation
at the Cherokee and Arapahoe Stations are considered open loop.
At the Valmont Station, a test program was implemented in October 1971* in
which one of the two scrubbing modules was modified so that limestone for pH
control could be added to the slurry feed. Each module receives 50 pet of
the incoming gas, or 25 pet of the total gas flow. Previous operation of the
scrubber system without pH control resulted in moderate to severe scaling, which
required a high degree of maintenance and thus, poor availability. The scaling
problem was a result, in part, of the high alkalinity of the Wyoming subbituminous
coal which removed U5 to 50 pet of the inlet sulfur dioxide which, in turn creates
supersaturated conditions with respect to calcium sulfate. Therefore, one
module was modified so that limestone could be added to the slurry feed in order
to adjust the pH to a range of 5.5 to 6.5. The level of suspended solids was
desired to be about 7 pet to provide seed crystals for the precipitation of
gypsum. Blowdown from the module is pumped to a lined settling pond. The
settling characteristics of the sludge is reported to be good and contains about
50 pet moisture. Overflow from the pond is pumped to a recycle tank where makeup
water is added at a rate of 70 gpm. The combined water is then used to slurry
297
-------
the limestone, wash demisters, and seal water for pumps. The only water loss
in the system is due to pond evaporation, evaporation in the module, and the
moisture in the sludge. The system is considered to be "closed loop."
Since the system has teen operated for only four months, the results are
considered to be preliminary in nature. Additional tests will be made in the
future. The module operating without limestone removes about U5 to 50 pet
of the incoming sulfur dioxide and has a particulate removal efficiency of
about 96 pet. The modified module, using limestone for pH control, removes
about 85 pet of the incoming sulfur dioxide and has a particulate removal
efficiency of about 96 pet, which meets particulate emission requirements of
0.02 gr/scf (wet). The state of sulfur oxidation is high, reported to be
nearly 100 pet. The degree of supersaturation is not known. Scaling in the
modified module is reported to be moderate, although additional tests are
required since some difficulty was experienced in maintaining the suspended
solids at 7 pet. The availability of the converted module was about 70 pet.
The particulate removal efficiency of the Cherokee and Arapahoe scrubber
systems are 95 to 98 pet, which meets the design goal of 95 to 98 pet. The
sulfur dioxide removal efficiency at the Cherokee Station, which burns Colorado
coal, is 15 to 20 pet. Sulfur dioxide removal at the Arapahoe Station, which
burns Wyoming coals, is U5 to 50 pet. The difference in sulfur dioxide removal
efficiencies can be attributed, in part, to different fly ash alkalinity.
The scrubber operating problems experienced to date are wear and periodic
replacement of the mobile bed packing, corrosive failure of the flue gas
reheaters, and scaling. At Valmont, the modified scrubber module experienced
moderate scaling on the module walls and certain portion of the first grid, whict
in turn caused pluggage of pump screens. Past experience at the other scrubber
modules using additives added for scale control, have not been successful,
although the tests were limited in nature. Scaling and plugging has occurred at
the wet/dry zone on the first stage grids, and on the reheaters. With the
exception of the Valmont experimental module, a higher than design blowdown is
practiced in an effort to control scaling and prevent pluggage.
Direct operating costs for Cherokee #3 is 0.50 mill/kwh, based on 75 pet
availability. Direct operating cost for the other stations are not available.
Scrubber operating requirements are electrical demand equal to k pet of the
power generated and also, steam for reheat. The water requirements are
approximately 3.3 to k.l acre-ft/MW/yr. Manpower requirements for scrubber
operations are not available. However, a high degree of maintenance is required.
Square Butte Electric Cooperative Pilot Plant
The Square Butte Electric Cooperative (SBEC) is currently constructing a
1*50 MW cyclone-fired generating unit requiring particulate and sulfur dioxide
abatement controls. The 1*50 MW unit is referred to as Center 2 and is being
constructed adjacent to the 238 MW Center 1 unit at the Milton R. Young Station.
Particulate control will be provided by electrostatic precipitators (ESP's) and
298
-------
sulfur dioxide control by wet scrubbers. SBEC and Sanderson & Porter, Inc.,
consulting engineers, selected Combustion Equipment Associates (CEA) and
Arthur D. Little, Inc. (ADL), to construct a 5,000 acfm (saturated) pilot plant
and the full-scale scrubber based partly on previous pilot plant experience
at Montana Power Company. The full-scale scrubber will utilize fly ash alkali
vith lime supplement. The design criteria and operating parameters of the full-
scale scrubber were determined in two months of testing conducted by CEA-ADL
in cooperation with SBEC, GFERC, and Minnesota Power and Light, on the 5,000
acfm pilot scrubber. The objectives of this program were:
1. To demonstrate that the stated performance guarantee
could be achieved for a maximum coal sulfur content of
1.3 pet.
2. To determine design conditions for L/G, percent suspended
solids in the recycle slurry, retention tank residence
time, and pH of the recycle slurry.
3. To determine supplemental lime requirments at the
maximum sulfur content.
^. To demonstrate the feasibility of using a flue gas bypass
for reheat.
A long-range testing program using the 5,000 acfm (saturated) pilot plant
will be conducted under a cooperative agreement between SBEC, Minnesota Power and
Light Company, Combustion Equipment Associates, and the Grand Forks Energy
Research Center. The objectives of the cooperative program are:
1. To determine whether sufficient alkali can be solubilized
from cyclone-fired fly ash to reduce sulfur dioxide in flue
gas below the level of State and Federal emission standards.
2. To determine the amount of additional alkali from lime which
may be required to supplement fly ash alkali to meet State
and Federal emission standards.
3. To determine the severity of calcium sulfite and calcium
sulfate scale formation under normal operating conditions
of the flue gas desulfurization pilot scrubber, and to
investigate chemical methods of minimizing the scale formation.
It. To establish that the pilot scrubber can be operated on a
closed-loop basis, and to determine the chemistry of the
closed-loop system.
5. To determine what effect fly ash-derived soluble salts in
the scrubber solution will have on the sulfur dioxide
removal efficiency.
299
-------
6. To determine and evaluate waste disposal of sulfate/sulfite
sludge and fly ash-derived soluble salts in sludge.
7. To conduct corrosion tests to determine the effects of
scrubber liquor on materials of construction to be used
for full-scale flue gas desulfurization processes.
8. To determine the mass balance of all input and output
materials, including selected trace elements and leachate
from sludge.
9. To evaluate the capital and operating costs of fly ash alkali
flue gas desulfurization for 100 MW, 500 MW, and 1000 MW
steam generator plants based on the technical and operating
data obtained from the pilot scrubber.
An additional phase of testing may be concerned with dilute sulfuric acid
scrubbing with fly ash neutralization, and sodium-magnesium and calcium double-
alkali-type scrubbing with fly ash neutralization.
A 5,000 acfm (saturated) pilot plant (about l.U MW equivalent) employing
spray nozzles was chosen to minimize the pressure drop (hence, energy usage)
across the absorption tower. The pilot plant (see Figure 10) has, essentially,
two liquid loops: the primary sulfur dioxide scrubber loops, and the 'mist
eliminator and wash tray loops.
The wash tray loop operates on clear liquor at approximately pH h and
consists of a wash tray above the absorber tower, a tray recycle tank, clarifier
and clarifier overflow tanks, and a demister. The wash tray, which is constructed
of 3l6L stainless steel, is designed to remove entrained droplets of scrubbing
liquor which could otherwise foul the demister. Liquid from the wash tray
drains to an 8 x 8-foot flakeglass-lined recycle tank which is then pumped back
to the wash tray or to an 8 x 8-foot flakeglass-lined clarifier. Overflow from
the clarifier is used to wash the bottom of the wash tray. Liquid from the
clarifier not used for washing is drained by gravity to a 6 x 6-foot flakeglass-
lined overflow tank. Makeup water from nearby Lake Nelson is added to the pilot
scrubber at the clarifier overflow tank at an average rate of l.U gpm (about 1.6
acre-ft/MW/yr) and the combined liquid is used to wash the polypropylene demister.
Underflow from the clarifier is pumped to a 3l6L stainless steel vacuum filter.
The scrubber loop operates on a slurry of alkali ash in recycled liquor at
a characteristic pH of h to 7- It consists of a 1*5 foot high by 3 1/2 foot
diameter flakeglass-lined absorber tower which contains six banks of 3l6L stainless
steel nozzles spraying scrubber liquid countercurrent to the gas flow. The
scrubber liquid drains to a 12 x 8-foot flakeglass-lined retention tank equipped
with a 3l6L stainless steel agitator. The retention tank liquid is pumped back to
the spray nozzles and also to an 8 x 8-foot flakeglass-lined thickener which is
used to control the level of suspended solids. Reducing the liquid flow from the
retention tank to the thickener will increase the level of suspended solids;
300
-------
To stub stack
Make up water
Filter cake
to disposal
Figure 10.- Flow diagram of 5,000 acfm saturated pilot plant scrubber, Square Butte
Electric Cooperative.
-------
increasing the flow rate will lower the level of suspended solids. Overflow from
the thickener drains by gravity to a 5 x 5-foot flakeglass-lined overflow tank
where it is mixed with liquid pimped from the vacuum filter. Thickener underflow
is pumped to the vacuum filter. The vacuum filter is operated only when the
concentration of suspended solids in the thickener underflow is above approximately
Uo pet. The liquid from the thickener overflow tank is pumped to a k x 5-foot
fly ash preparation tank with the excess liquid returning to the retention tank.
The fly ash slurry is pumped to an 8 x 8-foot feed tank for additional mixing
time, and then to the retention tank. Fly ash is stored in a 3 x 5-foot hopper
and fed to the preparation tank using a screw feeder at rates up to 8 lb/min. Lime
from a 3 x 2-foot storage hopper and screw feeder is used for pH control and is
fed directly into the retention tank without slaking. The total amount of
liquid in the entire pilot plant scrubber is about 25000 gallons. All pumps are
rubber lined. Liquid flows are measured by rotometer; liquid and gas temperatures
are measured by dial thermometers, and pressure drops are measured by manometers.
The pilot plant scrubber has the necessary equipment and controls to operate
over a wide range of variables. The solution pH can be varied from below pH it up
to pH 9; the retention tank residence time may be varied from k minutes to 16
minutes; the liquid to gas ratio may be varied from UO to 80; and a sulfur
dioxide injection system, can adjust the scrubber inlet sulfur dioxide to any
desired concentration. A duct equipped with an orifice and damper has been
installed between the inlet and outlet of the absorption tower and can be used to
bypass part of the hot inlet flue gas to mix with the cooler outlet flue gas
leaving the absorption tower. The mixing of flue gases in this manner is being
tested as a method for reheating to a temperature above the saturation point to
eliminate the possibility of stack gas rain.
A mobile trailer, constructed by KVB Industries according to specifications
supplied by the Grand Forks Energy Research Center, provided the capability of
continuously monitoring both the inlet and outlet flue gas for sulfur dioxide,
nitrogen oxides, carbon dioxide and oxygen. In addition to the gas monitoring
equipment, the trailer contains a chemistry laboratory to perform most analyses
of coal and scrubber liquor on site.
The preliminary results of the test program completed to date have been
favorable. The sulfur dioxide removal is required to comply with the Federal
emission standard of 1.2 Ib/MM Btu, which corresponds to approximately 535 ppm
S02 (dry). The sulfur dioxide removal efficiency was investigated as a function
of L/G, suspended solids, inlet sulfur, sulfur dioxide concentration, and fly
ash add rates. The ESP inlet fly ash particulate loading at the inlet to the
ESP on Center 1 ranges from 0.71 to 1.53 gr/scf and averages 1.13 gr/scf. The
three ash add rates investigated were equivalent to the average amount collected
by the ESP on Unit 2, the combined average amount collected by the ESP's on
Units 1 and 2, and the maximum amount .collected on Units 1 and 2. A typical
analysis of the fly ash is shown in Table 7.
302
-------
TABLE 7 A TYPICAL ANALYSIS OF THE FLY ASH PRODUCED
BY THE CYCLONE-FIRED CENTER UNIT NO. 1
Percent of ash,
as received
Loss on ignition at 800° C
Silica, Si02
Aluminum oxide , AlgO^
Titanium oxide , TiOg
Phosphorous pentoxide , P20j
Magnesium oxide , MgO
Potassium oxide, KgO
Sulfur trioxide, S03
2.2
29.8
12.7
10.6
0.5
0.3
25.7
U.5
2.2
2.0
6.U
TOTAL
Figure 11 illustrates the sulfur dioxide removal efficiency at the above
fly ash add rates at L/G ratios of 60 and 80. The dashed line corresponds to
the average fly ash production collected by both units, hereafter referred to
as the average ash add rate. The solid line corresponds to the maximum fly
ash production by both units, hereafter referred to as the maximum ash add rate.
A sulfur dioxide level of about 1100 ppm (dry) would be equivalent to about a
0.75 pet sulfur coal (HHV660J*). A level of 1900 ppm (dry) is equivalent to
about 1.3 pet sulfur in coal. The oulet sulfur dioxide represents the removal
for the total scrubber system, which includes the flue gas by-passed and used
for reheat. The total flue gas into the system was 6300 acfm, of which 1100 acfm
by-passed the absorber tower to provide reheat. The inlet gas temperature was
about 325° 7. The temperature of the saturated gas out of the absorber tower
was about 135° F. After mixing the by-pass gas, the temperature of the gas
to the stack was about 155° F. The flue gas reheat was tested as an alternative
to coil reheaters. No stack gas mist was observed to occur.
At a L/G of 60, an inlet level of about 1100 ppm S02 (dry) using fly ash at
the average add rate, the sulfur dioxide removal efficiency for the total
scrubber system was about 6l pet (absorber tower was about 70 pet). The total
fly ash alkali utilization, based on 25 pet CaO, k.12 pet MgO, and 1.67 pet
Na20, is about 63 pet. At the maximum ash add rate, the sulfur dioxide removal
is about 69 pet (absorber tower was about Bk pet), which represents a total
alkali utilization of 51.5 pet. Supplemental lime was not added and the
corresponding pH of the recycle slurry was about 3-9 at the average ash add rate
and U.9 at the maximum ash add rate.
At a L/G of 60, inlet level of about 1850 ppm S02 (dry), using the average
fly ash add rate with lime supplement, the total scrubber system removal efficiency
was about 70 pet (absorber tower removal of about 8^ pet). The supplemental
lime was added to maintain the pH at 6.0 to 7-0, and was chemically equivalent
from about 90 to 100 pet of the 1850 ppm inlet sulfur dioxide; the total alkali
303
-------
80O
I 1 1 1 1
1.2 Ib SOg/mmbtu
800 1,000 1,200 1,400 1,600 1,800 2,OOO
INLET S02 , Ppm (dry)
800
E
o.
. 600
M
O
UJ
O
400
200
1.2 Ib S02 / mmbtu
L/G = 60
800 1,000
1,200 1,400 1,600 1,800 2,OOO
INLET S02 , ppm (dry)
Figure II. - Sulfur dioxide removals in SBEC pilot plant tests
using fly ash alkali. Solid line represents average fly ash (1.13 gr/scf)
collected by ESP'S and dashed line represents maximum fly ash (1.53
gr/scf) collected by ESP'S.
304
-------
(fly ash alkali and lime supplement) was equivalent to about 1^0 pet of the
inlet sulfur dioxide. At the maximum ash add rate with lime supplement, the
sulfur dioxide removal remained at about 70 pet (absorber tower about Qh pet).
The supplemental lime was added to maintain the pH at 6.0 to 7.0 and was chemically
equivalent to about 68 pet of the inlet sulfur dioxide; the total alkali (fly
ash alkali and lime supplement) was equivalent to about lh2 pet of the inlet
sulfur dioxide. In a separate test using only lime chemically equivalent to
that of the total alkali, the removal efficiency increased to about 79 pet.
At an L/G of 80, inlet level of about 1100 ppm SOg (dry) using fly ash at
the average add rate, the sulfur dioxide removal efficiency for the total
scrubber system increased to about 81* pet (about 86 pet in the absorber tower).
The total fly ash alkali availability was increased to about 70 pet. The pH of
the recycle slurry was about 3.8. No supplemental lime was used. At the
maximum ash add rate with'supplemental lime, the total scrubber system sulfur
dioxide removal was about 8l pet (absorber tower was about 97 pet). The supplemental
lime was added to maintain the pH at 6.0 to 7-0, and was equivalent to about
36 pet of the inlet sulfur dioxide; the total alkali (fly ash alkali and
supplemental lime) was equivalent to about 162 pet of the inlet sulfur dioxide.
At an L/G of 80, inlet level of about 1850 ppm S02 (dry), the removal
efficiency for the total scrubber system was about 75 pet at the average ash add
rate (absorber tower was about 91 pet). Supplemental lime was added to maintain
the pH at 6.0 to 7.0, and was chemically equivalent to about 55 pet of the inlet
sulfur dioxide; the total alkali (fly ash alkali and supplemental lime) was
equivalent to 102 pet of the inlet sulfur dioxide. At the maximum add rate, the
removal efficiency for the total scrubber system was about 80 pet (absorber
tover about 97 pet), supplemental lime was added to maintain the pH at 6.0 to
7.0, and was chemically equivalent to about 53 pet of the inlet sulfur dioxide;
the total alkali was equivalent to about 128 pet of the inlet sulfur dioxide. A
test using only lime chemically equivalent to about 13^ pet of the inlet sulfur
dioxide, the high ash add rate with supplemental lime, resulted in a scrubber
system removal of about 80 pet (absorber tower about 97 pet).
The higher sulfur dioxide removals demonstrated in pilot plant tests were
obtained by adding lime at rates higher than intended for the H50 MW scrubber
unit, and these high rates may not be reproduced in practice on the commercial
scale scrubbers. The pilot test results do, however, demonstrate that the
scrubber design to be used on the full-scale unit is capable of meeting and
exceeding removals required to comply with the 1.2 Ib/MM Btu Federal emission
standard, and further, that required removals under normal conditions of coal
sulfur content can be achieved using fly ash alone without lime.
At the conclusion of the eight-week testing program, the scrubber system
was inspected for scale and it was reported to be light, with most deposits at
wet-dry interfaces. On the basis of previous pilot plant experience at Montana
Power Company, and limited confirmatory testing in this program, it was concluded
by CEA-ADL that a suspended solids level of 12 pet provides seed crystals
for precipitation of gypsum, and also provides scale control when recirculated to
305
-------
the spray nozzles. However, future reliability tests will be performed to
further investigate scaling and corrosion problems. The state of oxidation was
high, usually greater than 97 pet. The sludge was reported to have excellent
settling properties and cake from the vacuum filter contained ^0 to 50 pet
moisture.
Operational problems were not numerous and were concerned with an occasional
plugging of pipes and erosion of plastic valves and spray nozzles due to the high
level of suspended solids.
Burning coal with a sulfur content of about 0.75 pet, CEA-ADL recommends
the following full-scale scrubber operating conditions: L/G of 80, 12 pet
suspended solids, slurry recycle pH of 5. At the maximum fly ash add rate, no
additional alkali would be necessary. Fifteen percent of the incoming flue
gas will be by-passed and used for reheat.
Burning coal with the maximum sulfur content (l.3 pet), CEA-ADL recommends
the following full-scale operating conditions: A L/G of 80, 12 pet suspended
solids, a pH of 6.5 to 6.8, k tons per hour of supplemental lime at the maximum
fly ash add rate. Fifteen percent of the incoming flue gas will be by-passed
and used for reheat.
Capital and operating cost estimates on the i+50 MW full-scale scrubber were
presented at the Mid-Continent Area Power Pool (MAPP) Environmental Workshop,
November 18, 19751?. The capital investment estimates for the SBEC project are:
- Material and installation cost 35.5 $/KW
- Additional contracts and site
preparation cost 10.0 $/KW
- Escalation, interest during
construction, consultant fees,
Power Company overhead 20.0 $/KW
The estimated annual operating costs are:
Unit Cost ^/MM Btu $/yr
- Fly ash 0 00
- Lime (CaO) k2 $/ton 2.13 $ 555,5^0
- Waste disposal
transportation (30 TPH) 0.80 $/ton 0.99 258,^60
- Operating and main-
tenance staff
(20 men) 20,000
$/man 1.23 ^00,000
- Maintenance (k pet of
capital investment) 1.71 639,000
- Capital fixed charges 6.k\ 2,092,500
Total Annual Cost 12.5 $3,9UU,500
306
-------
If an equivalent lime scrubber were to be constructed, the following
operating costs would be affected: an increase of $1*50,000 per year to purchase
50 tons of lime per day, an increase of about $100,000 per year for offsite
disposal of fly ash, a decrease of about $50,000 per year since lower L/G could
be employed. The normal operating power requirement for the 1*50,000 KW fly ash
scrubbing system is reported to be 7922 KW. The maintenance and capital fixed
charge would be less since it is estimated that the capital investment for the
lime system is about 2.5 pet less than the fly ash system. When all economic
factors are taken into consideration, the Square Butte Electric Cooperative
expects the fly ash alkali scrubbing system to save I'M ,000 per year in operational
costs.
SUMMARY AND CONCLUSIONS OF WESTERN SCRUBBERS
Most future utility scrubbers for Western coals will be designed for the
primary purpose of sulfur dioxide removal and not for particulate removal, as
was true in the majority of past installations. The major reason is the
necessity for having flue gas desulfurization for most low sulfur Western coals
under the developing regulatory pattern in the West. There is also a trend toward
installing scrubbers in series with electrostatic precipitators on new plants
because of the appeal of eliminating the visible fly ash plume with a relatively
more reliable device, the ESP. If the scrubber can then be bypassed for short
periods for maintenance, a high plant availability can be maintained.
The three most significant factors to be considered in designing scrubbers
for Western coals are: l) the low concentration of sulfur dioxide to be removed,
2) the alkalinity of Western coal fly ashes, and 3) the tendency to operate at a
high state of oxidation, producing sulfate and not sulfite. It is significant
that all Western FGD installations, existing or planned, use throwaway type
processes. This is consistent with the remoteness of Western scrubbers from most
large markets for sulfur or gypsum, although sulfuric acid could possibly find
sizable markets in the West as an ingredient in fertilizer manufacture. Sulfur
dioxide emissions from low-sulfur Western coals can be brought into compliance
with the Federal emission standard by treating only a portion of the flue gas
and by-passing the remainder for stack gas reheat, as planned by Square Butte
Electric Cooperative. Cost is minimized by balancing the savings of treating
a smaller volume of gas against the cost of removing a higher percentage of the
sulfur dioxide from the fraction treated. Where local standards are more stringent,
the option of treating a partial flow ceases to exit.
At low concentrations of sulfur dioxide, it can be argued that gas film
diffusion should be the rate controlling step, since the equilibrium partial
pressure of sulfur dioxide for alkaline slurry is low and the capacity of the
slurry to absorb sulfur dioxide during passage through the scrubber is not taxed
if the amount absorbed is small. A sufficient L/G of course plays a part in
validating this argument. If gas diffusion does control, design should maximize
gas-liquid contact and residence time. Long residence time necessitates a large
volume; and good contact requires either multiple sprays or tower packing. If
scaling can be resolved, packing is probably the economical choice. If scale
has a tendency to form, the large empty volume with multiple sprays will be the
better option. The operating experience of the utilities cited in this paper is
that, if blowdown is sufficiently restricted, scale will indeed tend to form.
307
-------
The problem of scaling is inexorably tied in with the question of vhat
constitutes "closed loop" operation. The practical answer is that a system is
"closed" if no liquid blowdown is deliberately removed and disposed of. Some
inadvertent loss in sludge cannot be eliminated. Beyond this, pond evaporation
of a saturated scrubbing liquor does remove sulfate from the system by precipitation,
even though liquor may be returned from the pond. Since the sludge and pond
losses vill vary with design and climate, every system will be "closed" to a
different extent.
Control of scaling must be approached differently for low-sulfur Western
coals than for Eastern coals. A high state of oxidation and the slight control
that can be afforded by manipulating pH in a sulfate system offer little hope
that unsaturated operation is possible for Western coals during closed loop
operation. That this can be accomplished for high sulfur Eastern coals depends
on a low state of oxidation, with consequent precipitation of calcium sulfite and
coprecipitation of gypsum from a solution that is not saturated with calcium
sulfate'. These conditions do not appear to exist for Western coal operations.
Control of scale formation, in the authors' opinion, will depend most directly on
circulating a sufficiently high level of suspended solids, and operating at a
constant pH, whether high or low. This is supported by the 6 to 12 pet level of
recirculated suspended solids, and lime or limestone for pH control, currently
practiced by the utilities cited in this paper.
Depending on the cost of reagent and the properties of the waste products
produced, there may be more or less motivation to improve the utilization of
alkalinity in Western fly ashes in scrubbing systems. Laboratory tests at the
Grand Forks Energy Research Center have shown that utilization of fly ash alkali
improves significantly as the pH is dropped from 5 to 3. Other operating variables
would have to be changed along with the pH, including the flow circuit, gas-
liquid contact, L/G, and slurry reaction times.
In conclusion, involvement of alkaline ash in a scrubbing system implies a
new variable which must be controlled. Since the analyses of Western ashes
vary sufficiently, this effectively rules out one tailor-made solution for
every scrubber application. As fly ash varies, the characteristics of sludge
will vary, and therefore extensive studies on properties of sludge, including
leaching of major and trace elements, will be required. Experiences on operating
scrubbers indicate that work is needed on materials of construction, and on
component design to improve reliability. The solution to stack cleaning problems
will require continuing development effort as long as there is a conscious
desire for improvement.
308
-------
APPENDIX
Scrubber Design and Operation
Four Corners Plant
Arizona Public Service Company
LOCATION
1. Farmington, New Mexico.
2. Elevation is 5300 feet.
3. Atmospheric pressure is 12.1 psi.
k. Annual precipitation is 8 inches.
5. Water supply for the plant comes from Morgan Lake, a man-made reservoir filled
from the San Juan River.
SCRUBBER APPLICATION
1. Particulate removal, retrofit.
2. Boilers equipped with scrubbers.
- Two 175 MW Riley pc-fired boilers (Units 1 and 2).
- One 225 MW Foster Wheeler pc-fired boiler (Unit 3).
3. Service date: Units 1 and 2, December 1971; Unit 3, January 1972.
k. Fuel is New Mexico subbituminous coal from the Navajo mine.
- 8900 Btu/lb.
- 12 pet moisture.
- 0.68 pet sulfur.
- 22 pet ash.
- k pet CaO in ash.
5. Flue gas entering the scrubbers.
- Boilers 1 and 2.
- 8lU,000 acfm.
- 3^0° F.
- 650 ppm S02-
- 12 gr/scf particulate.
- Boiler 3.
- 1,030,000 acfm.
- Conditions are the same as on units 1 and 2.
7. The particulate removal goal was set by the project at 99-2 pet.
SCRUBBER DESCRIPTION
1. Two venturi scrubbers on each of boilers 1, 2, and 3.
,2. Vendor, the Chemico Air Pollution Control Company.
3. Capital cost is $30 million, or $52/kw.
1*. Operating costs are not available.
5. Materials of construction.
- Scrubbers are carbon steel with stainless steel or plastic lining.
- Outlet ducts were stainless steel lined, later lined with plastic over the
stainless steel.
- Liquid lines and pumps are rubber lined.
309
-------
- Process vessels are plastic lined.
- Reheaters were 3l6L stainless steel.
- Wet fans are inconel.
6. No "bypass.
T. Turndown is to approximately 50 pet of rated scrubber capacity.
8. Chevron mist eliminators have 6 stages.
9. Wet fan.
10. Reheaters that heated flue gas directly with steam coils failed because of
corrosion. The reheat units were removed about one year ago, and no reheat
has been used since. Indirect reheat by mixing with heated air is being
considered.
SCRUBBER OPERATING DATA
1. L/G is 8.5 gal/1000 acf, or 18 gal/1000 scf.
2. AP is 20 to 22 inches H20 across the venturi, 28 inches overall.
3. "Open loop." Total makeup water for the system is 1700 to 2000 gpm.
k. Gas residence time in the scrubber is not available.
5. Liquid delay time in the venturi recycle loop is about 2 minutes.
6. Liquid temperature leaving the scrubber is 120° F.
T. Solids recirculated has recently been increased from 2 pet to 6 pet.
8. pH of the recycle loop on the scrubber is 3.2 to 3-5. pH at the thickener
is U to 5 without lime added. A level of pH 7-5 at the thickener is to
be maintained with lime addition.
9. Scrubbing liquor analysis is not available.
10. State of oxidation is not available.
11. Degree of supersaturation is not available.
OPERATING REQUIREMENTS
1. Lime is added at the rate of 6 tons/day for pH control.
2. No dispersing agent is being used.
3. System makeup water requirements are about 3^00 acre ft/yr.
h. Power requirements.
- Electrical requirements are 3 to k pet of generating capacity.
5. Manpower.
- 8 operators.
- Maintenance personnel not available.
OPERATING RESULTS
1. Particulate removal meets the goal of 99•2 pet.
2. S02 removal is 30 to 35 pet without lime.
- No typical S02 removal has been determined with lime.
3. Availability overall is estimated at 80 pet.
it. Scaling has occurred on most wetted surfaces.
5. Methods used for scale control.
- The level of recirculated solids was recently increased from 2 to 6 pet.
- A lime system has recently been installed to maintain pH at 7-5 in the
thickener.
- The amount of blowdown used helps to control scaling.
310
-------
Problems.
- The principal problem is that scaling is not under control. The effects
of high pH in the thickener and 6 pet solids in the recirculation
scrubber liquor have not yet been assessed because of their recent
implementation, and a long-range testing program is planned.
- Corrosion with resulting leakage has occurred where coating on mild steel
have failed.
- Solids buildup has occurred in blowdown lines.
- Deterioration of stack linings.
Disposal of sludge.
- Sludge settles well without a flocculating agent. The sludge is concentrated
to 30 or 35 pet solids in the thickener underflow and is pumped after
some dilution with blowdown to decanting ponds. Ash at present is left
to accumulate in the ponds, but it may be dredged and returned to the mine.
Scrubber Design and Operation
Aurora Station
Minnesota Power and Light
LOCATION
1. Aurora, Minnesota.
2. Elevation 1500 feet.
3. Atmospheric pressure lU.l psi.
4. Annual rainfall approximately 25 inches.
SCRUBBER APPLICATION
1. Particulate removal, retrofit.
2. Two 58-MW pc-fired boilers.
3. Scrubber startup, June 1971-
1*. Fuel is Montana subbituminous coal from the Big Sky Mine.
- 8800 Btu/lb.
- 26.7 pet moisture.
- 1.37 pet sulfur (mine average).
- 9 pet ash.
- 9 to 13 pet CaO in ash.
5. Flue gas entering the scrubber on each boiler.
- 291,160 acfm.
- 3l+00 F.
- 800 ppm S02.
- 2.06 gr/scf.
d. Removal goals.
- Minnesota particulate standard is 0.6 Ib/MM Btu.
- Scrubber guarantee is 0.03 gr/scf or 0.078 Ib/MM Btu.
SCRUBBER DESCRIPTION
1. One Elbair spray-impingement scrubber for each of the two boilers.
2. One-stage of high pressure spray is directed concurrent with gas flow against
vertical rods to be atomized into fine droplets.
311
-------
3. Vendor, Krebs Engineers.
U. Size of each scrubber is nominally 60 MW or 300,000 acfm.
5. Capital cost is not available.
6. Operating cost is not available.
T. Materials of construction.
- Scrubber is 3l6 ELC stainless steel.
- Outlet ducts are flake-polyester coated carbon steel.
- Piping is fiberglass, or rubber lined.
- Pumps are rubber lined.
8. Ho bypass.
9. Turndown is 0 to 110 pet.
10. Demistor consists of one bank of vertical chevrons.
11. No reheat, wet fan.
SCRUBBER OPERATING DATA
1. L/G is 8.3 gal/1000 acf, or 13.3 gal/1000 scf.
2 AP is 2.5 inches of H20 across scrubber, k inches total.
3' Not closed loop, although scrubbing liquor is recycled from the ash pond
back to the scrubber. No clarifier is used. Makeup water is approximately
1200 gpm for each unit.
U. Gas residence time in the scrubber is approximately 3 seconds.
5. Liquid delay time in the recycle circuit is not available.
6. Solids circulated.
- 0.02 pet entering scrubber.
- 0.75 pet leaving scrubber.
7. pH estimated at 1*.5 in, U.1* out.
8. Scrubbing liquor analysis is not available.
9 State of oxidation of dissolved sulfur is high, estimted over 90 pet sulfate.
lo'. Saturation of scrubbing liquor does not occur at the level of blowdown used.
OPERATING REQUIREMENTS
1. No reagent is used.
2. Fcrubber water requirements are about 3500 acre ft/yr.
3. Power requirement.
- Electrical power is about 0.5 MW per unit, or 0.0 pet.
- No steam is used for reheat.
U. Manpower requirements are not available.
OPERATING RESULTS
1. Particulate removal is about 98 pet; exit dust loading is 0.0^ to 0.0^6 gr/scf.
2. S0? removal, occurring incidental to particulate removal by reaction with
the alkaline fly ash and by sulfate removal in blowdown, is typically 20 pet.
3. Availability. . .
- The Elbair scrubber, consisting of a stainless steel box containing hign
pressure spray nozzles that are removeable section by section for
maintenance, can remain on line without a bypass, even though the spray
system is not operating. However, effective operation requires a high
level of maintenance effort. A percentage availability in terms of
effective operation was not obtainable.
312
-------
Scaling and plugging.
- Scaling and plugging has not "been too severe, owing to the amount of
blowdown used.
Disposal of ash and spent scrubbing liquor.
- Slowdown, estimated to contain 1 pet solid fly ash, is sent an ash pond.
Overflow from the pond is neutralized with lime "before disposal.
Problems.
- This unit would suffer from the same scaling and plugging problems as the
Clay Boswell scrubber of similar design, described in the preceding
section, if the recycle loop were closed to a similar extent. As
operated, it is less troublesome.
- The mist carryover problem is less severe than at Clay Boswell owing to
operation at partial load with a resultant lower stack velocity.
Scrubber Design and Operation
Clay Boswell Plant
Minnesota Power and Light
LOCATION
1. Cohasset, Minnesota.
2. Elevation approximately TOO feet.
3. Atmospheric pressure lU.3 psi.
H. Annual rainfall approximately 25 inches.
SCRUBBER APPLICATION
1. Particulate removal for a new plant.
2. 350 MW Combustion Engineering pc-fired boiler.
3. Scrubber startup May 1973.
k. Fuel is Montana subbituminous coal from the Big Sky Mine,
- 8800 Btu/lb.
- 26.5 pet moisture.
- 1.37 pet sulfur (mine average)
- 9 pet ash.
- 9 to 13 pet CaO in the ash.
5. Flue gas entering the scrubber.
- 1,300,000 acfm.
- 251*0 F.
- 800 ppm S02-
- 3 gr/scf.
£. Removal goals.
- Minnesota particulate standard is 0.6 Ib/MM Btu.
- Scrubber guarantee is 0.03 gr/scf or 0.078 Ib/MM Btu.
313
-------
SCRUBBER DESCRIPTION
1. A single Elbair spray-impingement scrubber.
2. One stage of high pressure spray is directed concurrent with gas flov
against punch plate "baffles to be atomized into fine droplets.
3. Vendor, Erebs Engineers.
U. Size of scrubber is nominally 350 MW or 1.3 x 10° acfm.
5. Capital cost is not available.
6. Operating cost is not available.
7. Materials of construction.
- Scrubber is 3l6 LC stainless steel.
- Outlet ducts are flake-polyester coated carbon steel.
- Piping is fiberglass or rubber lined.
- Pumps are rubber lined.
8. No bypass.
9. Turndown is 0 to 110 pet.
10. Demistor consists of one bank of vertical chevrons.
11. No reheat, wet fan.
SCRUBBER OPERATING DATA
1. L/G — 8.3 gal/1000 acf, or 13.3 gal/1000 scf.
2. AP is 2.k inches 1^0 across scrubber, k inches total.
3. Not "closed loop" although scrubbing liquor is recycled from the clarifier
back to the scrubber. Makeup water is approximately 1^00 gpm. About
1*20 gpm of the makeup water compensates for clarifier underflow to the
fly ash pond. The remainder is clarifier overflow which is pumped to
the bottom ash pond. The clarifier overflow blowdown maintains the
sulfur level at about 3000 ppm.
U. Gas residence time in the scrubber is 3 seconds.
5. Liquid delay time in one clarifier is 2 hours, or U hours if two clarifiers
are on line.
6. Solids circulated.
- 0.02 pet entering scrubber.
- 0.75 pet leaving the scrubber.
7. pH 4.5 in, 4.0 out.
8. Scrubbing liquor analysis.
- Ca — 600 ppm.
- Mg — 200 ppm.
- Na — 15 ppm.
- SOlt — 2300 ppm.
9. State of oxidation of dissolved sulfur is high, estimated over 90 pet
sulfate.
10. Degree of supersaturation is not measured, but it is evident that it is
variable depending on the amount of soluble alkali in the ash, the amount
of S02 absorbed, and the amount of blowdown removed from the system.
OPERATING REQUIREMENTS
1. No reagent is regularly used.
2. Water requirements are about 2300 acre ft/yr.
314
-------
3. Power requirement.
- Electrical power is about 3 MW, or 0.86 pet of net generating capacity.
- No steam is used for reheat.
k. Manpower requirements are not available, but maintenance is known to be
very high because of a continuous schedule of cleaning on the mist
eliminators, which require sand blasting to clean. As of July 197U, 85
to 100 manhours were spent each week removing fly ash and calcium sulfate
scale deposits. In the past, the spray nozzles also required continuous
maintenance, however, with the increased blowdown now employed, the nozzles
require only periodic maintenance.
OPERATING RESULTS
1. Particulate removal is about 97 pet; exit dust loading is 0.08 pet
gr/scf.
2. S02 removal, occurring incidental to particulate removal by reaction with
the alkaline fly ash and by sulfate removal in blowdown, is typically
15 to 20 pet.
3. Availability.
- The Elbair scrubber, consisting of a stainless steel box containing
high pressure spray nozzles that are removable for maintenance section
by section, can remain on line without a bypass, even though the spray
system is not operating. However, effective operation requires a very
high level of maintenance effort. A percentage availability in terms
of effective operation was not obtainable.
k. Scaling and plugging.
- Scaling and plugging occur in nozzles, nozzle trees, strainers, on
punch plate baffles, in the wet-dry zone, and on the fan and mist eliminator;
and deposits fall into the drains at the bottom of the scrubber.
- Scaling is aggravated by any increase in CaO content in the ash of the
coal being burned, which causes increased SOg removal and an increase
in the level of Ca++ and SO^ ions in solution.
5. Measures used for scale control.
- Substantial amounts of blowdown are removed from the system to remain
below saturation with CaSOlj.
- A very high level of cleaning and maintenance is carried out continuously.
Nozzles and nozzle trees are removed and cleaned on a rotating schedule,
with all nozzles being cleaned once per week.
6. Disposal of ash and spent scrubbing liquor.
- Blowdown from two clarifiers, containing typically 5 or 6 pet fly ash,
is sent to an 80-acre ash pond. There is no net evaporation, but
rather an accumulation of about 10 inches of water per annum, z
7. Problems.
- Ho ultimate solution to the problem of disposing of sulfate-laden
blowdown water has been found. The two possible solutions would
involve either unrestricted discharge of diluted blowdown to streams
or closing the loop in the scrubber circuit to eliminate blowdown.
A break-through in methods of scale control would be required to
eliminate blowdown.
315
-------
- A stack mist problem results primarily from washing of the wet fan to
remove scale buildup. Attempts to use steam soot blowers in place
of washing and to apply non-stick coating have been unsuccessful.
Some improvement has been achieved by reducing the amount of fan
washing. Recent tests show favorable results by using low-velocity
stacks.
Because of the difficulties experienced with flue gas mist carryover, scale
deposits on the mist eliminators and ID fan, disposal of sulfate-laden
water and the high maintenance rate, the construction of similar scrubbers
for the two 70 MW boilers was not completed. In lieu of the particulate
scrubbers, ESP's are presently planned. Flue gas from all three units
would be discharged from one low velocity stack and the two 70 MW boilers
should provide reheat for flue gas exiting the 350 MW scrubber.
Scrubber Design and Operation
Dave Johnston Plant
Pacific Power and Light
LOCATION
1. Glenrock, Wyoming
2. Elevation approximately 5000 feet.
3. Atmospheric pressure 12.3 psi.
k. Annual precipitation is l*t inches.
5. Plant water supply is the North Platte River.
SCRUBBER APPLICATION
1. Particulate removal, retrofit.
2. One 330 MW Combustion Engineering pc-fired boiler (Unit no. k),
3. Three parallel scrubbers.
k. Scrubber startup was April 1972.
5. Fuel is Wyoming subbituminous coal from a captive mine. Typical analysis is:
- 7^30 Btu/lb.
- 26 pet moisture.
- 0.5 pet sulfur.
- 12 pet ash.
- 20 pet CaO in ash.
6. Flue gas entering scrubber.
- 1,500,000 acfm.
- 270° F.
- 500 ppm S02-
- 12 gr/scf (design).
- k gr/scf (actual).
7. Removal goal.
- 99-7 pet removal, or O.Oh gr/scf exit dust loading.
316
-------
SCRUBBER DESCRIPTION
1. Three identical scrubbers in parallel.
2. Venturi scrubbers.
3. Vendor, the Chemical Construction Company (Chemico).
U. Initial capital investment was $8 million, $2^/lew. Costs incurred
since startup have increased this amount significantly.
5. Operating cost is not available.
6. Materials of construction.
- Scrubbers, vessels, outlet duct, and stack are polyester lined steel.
- Piping and fan housing are rubber lined.
- Fan wheels are Inconel.
7- No bypass.
8. Turndown is to approximately 30 pet of rated scrubber capacity.
9. Chevron mist eliminator.
10. Wet fans, no reheat,
SCRUBBER OPERATING DATA
1. L/G is 13.3 gal/1000 acf, or 22 gal/1000 scf.
2. AP is 10 inches of H20 across the venturi, 15 inches total.
3. Intermittently "open loop." Normal operation is attempted at a makeup rate
of 500 gpm, which compensates for evaporation and loss in sludge. The
unit has been operated at times with 3000 gpm fresh water makeup to flush
out scale.
U. Gas residence time in the venturi section of the scrubber is estimated
at about 1 second.
5. Liquid exit temperature is 126° F.
6. Liquid delay time in the venturi recycle loop is 2 to 3 minutes.
7. Solids recirculated are 2 pet of scrubbing liquor.
8. pH leaving the scrubber is about 5-
Future tests will be conducted maintaining the pH at 5.5 to 6.0 using
hydrated lime.
9. Scrubbing liquor analysis is not available.
10. Degree of supersaturation is from 1.0 to 1.3.
OPERATING REQUIREMENTS
1. Lime is added for pH control.
2. Lignosulfonate is added to minimize scale at wet-dry zones.
3. Water requirements are approximately 800 acre-ft/yr in "closed loop" mode.
Actual requirements are greater because of occasional flushing.
k. Power requirements.
- Electrical power requirement is 7 to 8 MW, or 2.3 pet of
generating capacity.
- No steam is used for reheat.
5. Manpower requirements are not available.
317
-------
OPERATING RESULTS
1. Particulate removal meets the outlet grain loading goal of 0.1* gr/scf.
2. S02 removal (preliminary values).
- 35 to hO pet without lime.
- Lime addition results in modest increase in S02 removal. Exact value
has not teen determined.
3. Availability not available, but it is not considered adequate for a utility
power source.
U. Scaling and plugging.
- Solids buildup has occurred at the wet-dry interface.
- Hard gypsum scale has formed in the scrubber vessels and piping.
- Solids plug bleed and recycle lines.
5. Methods for controlling scaling and plugging.
- Lime for pH control has resulted in reduction but not elimination of
scaling.
- Lignosulfonate addition has resulted in a less adherent or friable wet-
dry buildup.
- Effects of hexamataphosphate on scaling and wet-dry buildup were determined
to be ineffective and its use discontinued.
- Continuous fan vash has essentially eliminated buildup on fans. The
wash water is composed of cooling water blowdown and service water.
- Fresh water washing has been required to flush ash and scale deposits
from the scrubber vessel.
6. Additional problems.
- Recycle pump erosion.
- "Silting" during shutdown.
7. Disposal of sludge.
- Bleed from the scrubber circuit is sent directly to two ash ponds, from
which overflow flows to a clear pond for recycle to the scrubber
circuit. Each ash pond is dredged once per year, and the solids
hauled out for landfill. Excess water resulting from periods of flushing
is discharged to the North Platte River under a variance from the State
of Wyoming.
Scrubber Design and Operation
Valmont, Cherokee, and Arapahoe Stations
Public Service Company of Colorado
NOTE: Public Service Company of Colorado has five TCA scrubbers containing
thirteen modules installed on five boilers at three stations. Because
of the close similarity between these installations, they will be discussed
collectively rather than individually.
318
-------
LOCATIONS
Valmopt Cherokee Arapahoe
Southwest Denver,
Boulder. Colo. Horth Dearer, Colo. Colo.
Elevation, ft 5300 Bst. 5200 Est. 5600
Atm. P, psi 12.1 12.1 12.1
Annual rainfall, in. Ik lit -^
Plant water supply Hillcrest Lake South Platte River South Platte River
SCRUBBER APPLICATION
1. All units are retrofits for particulate removal; Valmont #5 has one module
modified to study SOg removal.
2. Boilers equipped with scrubbers (all pc-fired).
3. Valmont #5, 196 MW, 2 scrubber modules, November 1971.
Cherokee #1, 115 MW, 1 scrubber vessel, 2 modules, June 1973.
Cherokee #3, 170 MW, 1 scrubber vessel, 3 modules, Bbvember 1972.
Cherokee A, 375 MW, 1 scrubber vessel, k modules, July 197^.
Arapahoe #U, 112 MW, 1 scrubber vessel, 2 modules, September 1973.
It. Arrangements of particulate cleaning equipment.
- At Valmont, flue gas from a mechanical collector is split into two parallel
streams, with 50 pet sent to the scrubber and 50 pet to an electrostatic
precipitator (ESP).
- All other units have a mechanical collector, an ESP, and scrubber(s) in
series, with all flue gas entering the scrubber(s).
5. Coal burned at Valmont and Arapahoe is Wyoming subbituminous.
- 8300 Btu/lb.
- 29 pet moisture.
- 0.6 pet sulfur.
- 5.2 pet ash.
- 20 pet CaO in ash.
6. Coal burned at Cherokee is Colorado bituminous coal.
- 11,000 Btu/lb.
- 9.8 pet moisture.
- 0.7 pet sulfur.
- 9-^ pet ash.
- 5 pet CaO in ash.
7. Flue gas entering the stack gas cleaning train.
Valjmont #5 Cherokee #1 Cherokee #3 Cherokee #U Arapahoe
#1*
Flow, acfm H63,000 520,000 610,000 1,520,000 520,000
T» °F 260 295 280 275 300
S02, ppm (est.) 500 500 500 500 500
gr/scfd 0.8 0.8 O.k 0.7 0.8
8. Removal goals for particulate.
- The applicable Colorado State Standard is 0.1 Ib/MM Btu, or about 0.05
gr/scf .
- The company's desire for clean stacks requires a goal of 0.02 gr/scf.
319
-------
SCRUBBER DESCRIPTION
1. All units are Turbulent Contact Absorbers, consisting of three stages of
motile' packing, or "ping pong balls," with spray directed downward through
the balls and gas passing countercurrent upward.
2. Vendor: Air Correction Division, Universal Oil Products Company.
3. Capital cost:
Valmont #5 Cherokee #1 Cherokee #3 Cherokee #k Arapahoe #h
#3,600,000 $3,810,000 $1*, 1*00,000 $12,700,000 $1*, 560,000
$32/kw $33/kw $29/kw $33/kv $1*1/kw
k. Operating costs for Cherokee #3 is 0.50 mills/kwh. Other operating costs
are not available.
5- Materials of construction (typical).
- Scrubbers are rubber-lined carbon steel with stainless steel grids.
- Exit ducts are mild steel with stainless steel and/or fabric expansion joints.
- Slurry piping is rubber lined.
- Pumps are rubber lined.
6. Bypasses on all units.
7. Typical turndovn is 1*7 to 105 pet.
8. Mist eliminators have 2 stages, 7 passes.
9. Demisters for Cherokee #1 and #3 are fiberglass reinforced plastic.
Demisters for Cherokee #1*, Arapahoe #1* and Valmont #5 are stainless steel.
10. All units except Cherokee #1+ reheat the flue gas directly with steam coils;
Cherokee #1* uses externally-heated air.
11. All units have dry fans that are forced-draft with respect to the scrubber.
SCRUBBER OPERATING DATA FOR CLOSED LOOP MODULE AT VALMONT #5
1. L/G of about 50.
2. AP is approximately 10 to 15 inches of J^O.
- Three stages of mobile packing (8-12 inches per stage).
- Stainless steel mist eliminators.
- Direct steam coil reheater, 1 1/2 to k inches.
- Transition ductwork, 1/2 inch.
- Total, 16 1/2 inches maximum.
3. Makeup water is 70 gpm. The water analysis is as follows:
- SOi, — 45 ppm.
- Ca — 35 ppm.
- Cl — 10 ppm.
h. Gas residence time is not known.
5. pH entering is maintained at 5-5 to 6.5 by limestone addition.
6. State of oxidation of dissolved sulfur is nearly 100 pet sulfate.
7. Suspended solids maintained at 7 pet.
8. Liquid temperature is 110° F.
9. Degree of supersaturation is not known.
320
-------
SCRUBBER OPERATING DATA FOR OPEN LOOP SYSTEMS
1. L/G is typically 5^-59-
2. AP is approximately 10 inches to 15 inches of ^0 depending on design
and operating conditions.
- 3 stages of mobile packing — 8-12 inches.
- Mist eliminators.
- Fiberglass reinforced plastic, 1 1/2 to 3 inches.
- Stainless steel — 1/2 inch.
- Reheat.
- Direct steam coil reheater, 1 1/2 to k inches.
- Hot air — negligible.
- Transition ductwork — 1/2 inch
- Total 16 1/2 inches maximum
3. "Open loop." Amounts of makeup water from cooling tower blowdown are as
follows:
Cherokee #1
203 gpm
Cherokee #3 Cherokee #U
380 gpm jkh gpm
Arapahoe #k
203 gpm
Valmont #5
(unmodified
module )
210 gpm
k. Gas residence time in the scrubber is 3.8 to 5 seconds.
5. Liquid temperature leaving the scrubber is 110° F.
6. Liquid holdup time in the recycle circuit is very short, estimated at
ten seconds.
7. pH is T to 9 entering; 2.8 to 3 leaving the scrubber.
8. A scrubber liquor analysis for Valmont (not known to be representative).
- Ca — 590 ppm.
- Mg — 350
- Na — 1
- SO^ — 10,000
9. State of oxidation of dissolved sulfur is not available.
10. Degree of super saturation is not available.
OPERATING REQUIREMENTS
1. No lime or other reagent or additive is normally used.
2. Water requirements in acre-ft/yr (approximate).
Valmont #5
3^0
Cherokee #1
327
Cherokee 13
612
Cherokee #h
1200
Arapahoe tik
327
3. Power requirements.
Electric: Power:
Valmont #5 Cherokee
Cherokee #3 Cherokee #k Arapahoe #k
5.3 MW
5.^ %
MW
*. 5
6.k m
3.8 %
it.U m
3.8 %
5.2 m
U.6 %
321
-------
Steam for reheat:
Valmont #5 Cherokee »1 Cherokee #3 Cherokee #k Arapahoe #h
Ib/hr: 50,000 50,000 1»1,200 135,000 60,000
°F: TOO teO 715 ^ 360
psi: 1*90 300 300 1,975 150
U. Manpower for scrubber operation is not identified "by the company as a
separate category from plant oeprators. However, it is reported that a
high degree of maintenance is required.
OPERATING RESULTS
1. Particulate removal achieves an outlet grain loading of 0.02 gr/scf.
2. S02 removal.
- ^5 to 50 pet at Valmont and Arapahoe, burning Wyoming coal,
- 15 to 20 pet at Cherokee, burning Colorado coal.
- 85 pet at Valmont using modified module (not used continuously).
3. Availability.
- Valmont #5 — 55 pet for k6 months of operation.
- Cherokee #1 — 53 pet for 23 months of operation.
- Cherokee #3 — 66 pet for 35 months of operation.
- Cherokee #U — 82 pet for 10 months of operation.
- Arapahoe #U — 8U pet for 25 months of operation.
^. Problems.
- Scaling and plugging has occurred at:
- The wet/dry zone.
- The first stage grid.
- The reheater.
- The mist eliminators.
- Corrosion has caused major failures of reheaters at the Cherokee Station.
- Wear on the balls used for packing requires replacement after 6000 hours
or less.
- The cost of replacing the balls are 0.06 $/ball. The number of balls
required is as follows:
- Valmont #5 — 870,000 balls.
- Cherokee #1 — 980,000 balls.
- Cherokee #3 — 1,180,000 balls.
- Cherokee #U — 3,070,000 balls.
- Arapahoe #k — 980,000 balls.
5. Measures for control of scaling.
- Additives tried have not worked, including phosphated esters. The studies
were limited in that they were of relatively short duration.
- Slowdown must be maintained at an adequately high level, but otherwise
no specific methods are being used.
- One of two modules at the Valmont Station has been modified to use
limestone for pH control. The preliminary results are favorable:
however, additional tests are planned for the future.
322
-------
6. Sludge disposal.
- At the Arapahoe Station slurry blowdown is pumped to an ash settling
pond after adjusting the pH to 6.0 - 9.0 with lime. The settled fly
ash is dredged periodically and used for landfill. Clear effluent
from the ponds is discharged under permit from the State of Colorado,
to the South Platte River at the Cherokee and Arapahoe Stations, and
to the cooling pond at the Valmont Station. At the Cherokee Station
only slurry blowdown is neutralized and clarified and then mixed with
ash pond overflow for discharge to the river, owing to the heavy
loading on the ponds caused by discharge from the three scrubber-equipped
boilers at this station.
iCKNOWLEDGMENTS
Information contained in this paper was obtained from published sources
and, from inquiries directed to'utilities operating wet scrubbers on
boilers burning Western coals and to scrubber vendors. Persons who have
contributed information include Dr. Fred Murad of Combustion Equipment Associates,
Mr. Ken Vig of Minnkota Power Cooperative, Mr. Dennis Van Tassel and
Ifr. ELdon Kilpatrick of Minnesota Power and Light, Mr. Phil Richmond of Square
Birtte Electric Cooperative, Mr. George Greene and Mr. Steven Goering of Public
Service of Colorado, Mr. Thomas Ashton of Pacific Power and Light Company,
Br. Carlton Grimm of Montana Power Company, Mr. John Noer of Northern States
Bower, Mr. Walter Ekstrom and Mr. Aubrey Parsons of Arizona Public Service. The
Contributions of the above and others who have freely exchanged information in
numerous past contacts concerning flue gas desulfurization are gratefully
acknowledged.
BH'ERENCES
1. U.S. Bureau of Mines, Division of Fossil Fuels. Coal—Bituminous and
Lignite in 1973. Mineral Industry Surveys, January k, 1975, p 5.
2. Nielsen, G.F. Coal Mine Development Survey. Coal Age. V. 80, February
1975, pp 130-139.
3. Energy Research and Development Administration. Open file report. Survey
of Coal and Ash Composition and Characteristics of Western Coals and
Lignite. Grand Forks, ND, 1975.
.\. Gronhovd, G.H., et al. Some Studies on Stack Emissions from Lignite Fired
Power Plants. BuMines 1C 8650, 197'k, pp 103, 133.
5. Facts Sheet. Four Corners Powerplant, Farmington, NM, April 1973.
6. Quig, R.H. Chemico Experience for 50^ Emission Control on Coal-Fired
Boilers. Presented at the Coal and the Environment Technical Conference,
Louisville, KY. October 23, 197^.
-------
7. Borgwardt, R.H. EPA/RTP Pilot Studies Related to Unsaturated Operation on
Lime and Limestone Scrubbers.
8. Tufte, P.H., et al. Pilot Plant Scrubber Tests to Remove S02 Using
Soluble Alkali in Western Coal Ash. BuMines 1C 8650, 197^, PP 103-133.
9 Sondreal, E.A. , et al. Wet Scrubbing of S02 with Alkali in Western Coal
Ash. Paper No. 7^-272, 67th Annual Meeting of the Air Pollution Control
Association, June 9-13, 197^, 31 pp.
10. Sondreal, E.A., and P.H. Tufte. Scrubber Developments in the West. Presented
at the Lignite Symposium, Grand Forks, North Dakota, May 1^-15, 1975-
11. LaMantia, C., et al. EPA-ADL Dual Alkali Program Interim Results. Presented
at EPA Symposium on Flue Gas Desulfurization, Atlanta, Georgia, November
12. Kilpatrick, E.R., and H.E. Bacon. Experience with a Flue Gas Scrubber on
Boilers Burning Subbituminous Coal. American Society of Mechanical
Engineers Winter Annual Meeting, New York, NY, November 1971*. Paper No.
71+-WA/APC-3.
13. LaMantia, C.R., and I. A. Raben. Some Alternatives for S02 Control. Presented
at Coal and the Environment, Technical Conference sponsored by the
National Coal Association, October 22-21* , 197)1.
Ik. Noer, J.A., et al. Results of a Prototype Scrubber Program for the
Sherburne County Generating Plant. Presented at the IEEE-ASME Joint
Power Generation Conference, Miami Beach, Florida, September 15-19,
15. Ashton, T.M. Operating Experience Report, Flue Gas Scrubbing System,
Dave Johnston Steam-Electric Plant Unit h, Pacific Power and Light
Company, presented at the American Society of Mechanical Engineers
National Symposium, Philadelphia, PA., April 1973.
16. Green, G.P. Operating Experience with Particulate Control Devices.
Presented at the American Society of Mechanical Engineers National Symposium,
Philadelphia, PA, April 1973.
17. Murad, F.Y., et al. Boiler Flue Gas Desulfurization by Fly Ash Alkali.
Presented at Mid-Continent Area Power Pool (MAPP) Environmental Workshop,
Minneapolis, MN, November 18, 1975-
324
-------
RESULTS OF THE 170 MW TEST MODULES PROGRAM
MOHAVE GENERATING STATION
SOUTHERN CALIFORNIA EDISON COMPANY
Alexander Weir, Jr., Lawrence T. Papay, Dale G. Jones,
John M. Johnson, and William C. Martin
Southern California Edison Company
P. 0. Box 800
2244 Walnut Grove Avenue
Rosemead, California 91770
ABSTRACT
This paper summarizes the performance of three different types
of scrubbers tested with lime and limestone reagents at the Mohave
Generating Station in South Point, Nevada. Each scrubber was designed
to treat 450,000 SCFM of flue gas (170 megawatt equivalent) and was
larger than any other single scrubbing module which has been operated
in the world today. The Horizontal Module (a horizontal cross flow
spray scrubber) was operated from January 16, 1974 to February 9, 1975.
The Vertical Module was tested in two modes; first as a Turbulent
Contact Absorber (TCA) from November 2, 1974 to April 30, 1975 when
the thermoplastic "ping pong balls" were removed and second as a Polygrid
Packed Absorber (PPA), with a plastic "eggcrate" packing with testing
continuing to July 2, 1975.
The effects of recirculating slurry flow rate, flue gas flow rate
(turndown) and number of scrubbing stages on S0_ and particulate removal
are presented. Performance of two types of flue gas reheaters and two
types of mist eliminators are discussed. Power requirements, reagent
utilization factors, scrubbing system, availability records and closed
loop water system operations are described for three types of scrubbing
systems tested at Mohave.
325
-------
FOREWORD
The Test Modules Program was a joint venture of the Navajo and
Mohave Power Project participants who are listed below:
Salt River Project Agriculture Improvement and Power District
Arizona Public Service Company
Department of Water and Power of the City of Los Angeles
Nevada Power Company
Tucson Gas and Electric Company
Bureau of Reclamation of the U.S. Department of the Interior
Southern California Edison Company
Funding for this program was provided by the participants in
accordance with their respective megawatt entitlements in the Navajo
and Mohave Power Projects. Southern California Edison Company was the
project manager of the Test Modules Program.
The conclusions presented in this paper represent the personal
opinions of the authors and are not intended to represent the opinions
or position of any of the project participants.
326
-------
INTRODUCTION
At the second EPA Flue Gas Desulfurization Symposium in New Orleans
in 1972, Dr. Jerry Shapiro1 of the Bechtel Power Corporation presented
the plans of the Navajo and Mohave participants to test, on a pilot plant
scale, four different types of scrubbers with three different reagents
(lime, limestone, and soda ash). The results of these tests, as well
as tests of four additional scrubbers with one additional reagent (ammonia),
were presented at the third Flue Gas Desulfurization Symposium in New
Orleans in May 1973. At that time it was indicated that contracts had
been let in December of 1972 to construct two 170 MW Test Modules at the
Mohave Generating Station. Both of these scrubbers were larger than any
other single scrubbing module which had been previously operated in the
United States, or in the world. With a rated capacity of 450,000 SCFM,
that statement- is still true today. Testing of the Horizontal Module
was initiated January 16, 1974 and the results of the first nine months
of operation were presented at the fourth Flue Gas Desulfurization Sym-
posium in Atlanta on November 4, 1974. Testing of this module was com-
pleted February 9, 1975 after 5927 hours of operation and it has since
been dismantled and reassembled at the Four Corners Generating Station.
Start-up of the Vertical Module was initiated on schedule January 1,
1974, but on January 24, 1974 a disasterous fire burned most of the
chlorobutyl rubber lining and start-up was not resumed until October 1,
1974.3 Testing of the Vertical Module was initiated November 1, 1974,
and was completed July 1, 1975 after 3131 hours of test. Two different
configurations of the Vertical Module were tested. The first configura-
tion was the TCA (Turbulent Contact Absorber) configuration in which four
Stages and later three stages of thermoplastic rubber balls were used.
Later, the balls were removed and replaced with a plastic "eggcrate"
type packing referred to in this paper as the PPA (Polygrid Packed Absorber)
configuration. The Vertical Module is presently shut down in a cold
standby condition at the Mohave Generating Station.
It is the purpose of this paper to present the results of testing
of these three types of scrubbers at the 170 MW scale, as well as results
of testing of two types of reheaters, and two types of demisters. Not
included in this paper, for lack of time, were the results of testing
various scrubber lining materials. Six types of slurry pumps, as well
as rotary filters, centrifuges and two commercially available methods
of sludge fixation (as well as sludge ponding) were also evaluated in
this program but these results also will not be presented in this paper.
Since another session of this symposium is directed to slurry and sludge
handling, EPA has requested that we not discuss this aspect of a total
scrubbing system in this paper.
327
-------
TEST MODULE DESCRIPTIONS
170 MW Vertical TCA and PPA
A ground-level photograph of the Vertical Module is shown in
Figure 1, where the flue gas enters the scrubbing chamber at the lower
left and exits from the top of the scrubber. Although this photograph
does not show the front end limestone preparation area, back-end thickener
nor sludge dewatering equipment, it does indicate the height to the
top of the outlet ductwork, which is about 145 feet above grade.
The ductwork shown at the left side of Figure 1 supplies flue gas
to the.scrubber from the down stream side of the Unit 1 electrostatic
precipitator. A 5,500 HP booster fan forces the flue gas through
the inlet ducting, scrubber packing and mist eliminator section into
the outlet ducting where it is heated by direct-contact steam coils
and is finally returned to the ductwork leading to the 500 foot stack.
The majority (2,342 hours) of the Vertical Module test program
was conducted with the Turbulent Contacting Absorber (TCA) configuration
shown in Figure 2. The scrubbing chamber dimensions are 18 feet wide
by 40 feet long and 38 feet high. The TCA configuration consisted
of four stages of thermoplastic rubber balls supported on stainless
steel grids at four foot intervals. The balls were contained in compart-
ments, with 15 compartments at each of the tour levels. Although various
levels of ball depths were tested, the compartments were initially
filled to the one-foot level with approximately 1,600,000 balls. The
TCA configuration was tested with both three and four stages of balls
and at ball depths of 6, 10 and 12 inches in the four stage configuration.
During that last phase of the program, the TCA configuration
was replaced with a Polygrid Packed Absorber (PPA), as shown in Figure
3. The Vertical Module was operated for 789 hours in the PPA configur-
ation. The Polygrid packing consisted of plastic grids in an "eggcrate"
configuration where each grid layer was 1 1/4" thick with 2" square
openings. The grid layers were stacked to a depth of approximately
17 inches in each stage.
The Vertical Module recirculating slurry flow rate was normally
16,000 gpm in the TCA configuration and 27,000 gpm in the PPA configur-
ation. A small quantity of the recirculating slurry was discharged
to a dewatering complex to extract solids from the scrubber slurry.
Water reclaimed in the dewatering process was returned to the limestone
slurry mix tank. The dewatering complex consisted of a thickener tank,
with further dewatering provided by either a centrifuge or vacuum drum
filters.
170 Horizontal Module
The photograph shown in Figure 4 shows the size and general
arrangement of the basic scrubber shell, ductwork and major components
of the Horizontal Module. The horizontal end-to-end length including
the booster fan, demister section and ductwork is 150 feet long. The
width varies from 28 to 32 feet and the top of the scrubber shell
is 33 feet above grade. The ductwork shown on the right side of Figure
328
-------
••
'• '
'
FIGURE1 - 170 MW VERTICAL MODULE
-------
Figure 2
I70MW VERTICAL MODULE (4STAGE TCA)
SLURRY SPRAYS
FLUE GAS
FROM FAN
SUMP CHAMBER
SCRUBBED FLUE GAS
TO DEMISTER
A
THERMO PLASTIC
RUBBER SPHERES
"QUIESCENT" SPHERE
DEPTH OF 6" PER
STAGE
330
-------
Figure 3
I70MW VERTICAL MODULE
( 3 STAGE POLYGRID PACKED ABSORBER )
SCRUBBED FLUE GAS
TO DEMISTER
SLURRY SPRAYS
17 PACKING PER
STAGE
SUMP CHAMBER
-------
r.i
•<
< <
FIGURE 4 - 170 MW HORIZONTAL MODULE
-------
4 supplies flue gas to the scrubber from the downstream side of the
Unit 2 electrostatic precipitator. A 1,750 HP booster fan forces the
flue gas through the scrubber shell and demister section into the outlet
ducting where it is heated by injecting hot ambient air and is finally
returned to the precipitator outlet duct upstream from the 500 foot
stack.
The four-stage scrubbing chamber of the Horizontal Module is
shown in Figure 5 and is 48 feet long, 28 feet wide and 15 feet high.
The Horizontal Module did not contain packing, but consisted of four
stages of cross flow spray. The slurry was cycled through the scrubber
in a countercurrent manner. That is, the fresh lime slurry from the
mix tank was first sprayed across the flue gas at the fourth stage,
or discharge end of the scrubbing chamber. The same liquid was successively
collected and pumped to the third, second, and first stages and success-
ively depleted of alkalinity. By the time the slurry reached the first
stage collection hopper for return to the lime mix tank, it was almost
completely depleted of any excess alkalinity.
The sprays were discharged from a row of 36 externally-mounted
nozzles at each stage. The recirculating slurry flow rate was normally
9,000 gpm, but the slurry was mechanically pumped four times per circuit
for a total installed pump capacity of 36,000 gpm. As in the Vertical
Module, a small quantity of recirculating slurry was discharged to
adewatering complex to extract solids from the scrubbing system.
Water reclaimed in the dewatering process was returned to the scrubber
slurry lime mix tank. However, in the case of the Horizontal Module,
the dewatering complex consisted of a thickener tank and a sludge
disposal pond.
TEST CONDITIONS
Both test modules were tested under a variety of configurations
and conditions to determine the significant parameters affecting SO2
and particulate removal, operating simplicity and system chemistry.
The test program for each module was divided into test blocks and
were designed with specific values for selected variables. This approach
was adopted to simplify data analysis and permit the effect of
each variable to be determined independently. The variables considered
aost relevant are listed below:
Type of reagent
Demistet wash procedure
Number of scrubbing stages
Percent cooling tower blowdown in makeup water
Spray nozzle pressure
Flue gas flow rate
Circulating liquid flow rate
Inlet S02 concentration
Inlet particulate grain loading
pH of the scrubber slurry
Method of water balance control
Percent solids in scrubber slurry
Percent solids in reagent feed slurry
-------
FIGURE 5
170 MW HORIZONTAL MODULE
(4 STAGE)
FLUE GAS
/ FROM FAN
SCRUBBER
FLUE GAS
TO DEMISTER
-------
TEST PERFORMANCE RESULTS
S02 Removal
«l,,rrv ^mparifon °f S°2 removal efficiency as a function of recirculating
slurry flow rate for the Horizontal and Vertical Modules is shown in
Figure 6. Note that with low sulfur Western coal fired at the Mohave
Generating Station, the S02 concentration in the stack gas and at
the scrubber inlet averaged only 200 ppm. At this low inlet S02 concen-
tration, bOz removal above 95% was achieved with all three scrubbers.
For example, the recirculating liquid flow rate required for 96% SO2
removal was 9,500 gpm with the Horizontal Module 18,500 gpm with the
Vertical Module in the TCA mode and 27,800 gpm with the Vertical Module
in the PPA mode.
Effect of Staging on SO2 Removal
The percentage S02 removal in the Horizontal Module as a function
of the number of scrubbing stages is shown in Figure 7. The first
stage removes 50% of the SO2. The second stage removes 64% of the
remaining S02 for a total S02 removal with two stages of 82%. Each
successive stage removes some of the remaining S02 and although the
total percentage removal increases, the total weight of S02 removed
in each additional stage decreases. The Horizontal Module design allows
a selection of the appropriate number of scrubbing stages to obtain the
desired degree of S02 removal. Since the Horizontal Module can continue
to operate with the loss of all but one stage, it was relatively easy
to obtain the effect of staging by selectively shutting off stages.
It was more difficult to obtain this type of data with the Vertical
Module since shutdown was required to remove the packing. However,
the Vertical Module was operated in the TCA mode with 3 and 4 stages
and in the PPA mode with 2 and 3 stages. The results of the effect
of staging on S02 removal were in general agreement with those obtained
with the Horizontal Module.
Effect of Turndown on S02 Removal
All three scrubbers showed a slightly increased degree of S02
removal when the gas flow rate was decreased from 450,000 SCFM and
the circulating liquor flow rate was allowed to remain constant. It
should be recognized that the SO2 concentrations at the exit of the
scrubber were extremely low and wet chemical analysis, rather than
instrument readings, were used. The measured exit valves were corrected
(i.e. increased) for the dilution effects of water vapor, and in the
case of the Horizontal Module for the reheat air so that they were
comparable to the inlet values. The effect of turndown on exit S02
concentration is shown in tabular form in Figure 8.
Particulate Removal and Turndown
At the design operating condition of 450,000 SCFM and 0.10 gr/SCF
inlet particulate grain loading, both the Horizontal and Vertical
Modules achieved a relatively high degree of particulate removal.
This is shown in Figure y, where the Vertical TCA achieved 93% removal,
the Horizontal Module ahieved 92.5% removal and the Vertical PPA
-------
FIGURE 6
EFFECT OF CIRCULATING LIQUOR FLOW RATE
ON S02 REMOVAL AT CONSTANT GAS FLOW
G=450,000 SCFM
100-
99-
98-
97-
TCA 4 STAGES
LIMESTONE
UJ
cc
CJ
UJ
UJ
Q_
UJ
95-
94-
93-
HORIZONTAL
4 STAGES
LIME
PPA 3 STAGES
LIMESTONE
92
91-
90-
10
15
20
25
30
CIRCULATING LIQUOR FLOW RATE
(1000 GPM) 336
-------
100
FIGURE 7
170 MW HORIZONTAL
S02 REMOVAL VS. NUMBER OF STAGES
so-
il!
8
I!
40-
CONDITIONS
GAS FLJOW= 450,000 SCFM
L/G = 206PM/IOOOSCFM
INLET S02= 220 ppm LIME
20
NUMBER OF STAGES
337
-------
FIGURE 8
EFFECT OF TURNDOWN ON S02 REMOVAL
00
NUMBER OF STAGES
REAGENT
CIRCULATING LIQUOR
FLOW RATE (GPM)
INLET S02 CONCENTRATION
(ppm)
EXIT S02 CONCENTRATION
(ppm)
at 450,000 SCFM INLET
at 300,000 SCFM INLET
HORIZONTAL
4
LIME
9,500
220
4
2
VERTICAL
TCA
18,000
220
II
9
VERTICAL
PPA
LIMESTONE LIMESTONE
27,000
220
8
3
-------
100-
FIGURE 9
EFFECT OF TURNDOWN ON ARTICULATE REMOVAL
INLET GRAIN LOADING=0.10 GRAINS/SCF
95-
HORIZONTAL CIRCULATING
LIQUOR = 9000 6PM
4 STAGES
-------
achieved 91.5% removal. Note that this is a relatively high weight per-
centage removal for the fine particles remaining in the flue gas after
passing through an upstream electrostatic precipitator.
Note that the particulate removal efficiency of the Vertical
TCA and PPA decreases as the flue gas flow rate is reduced, while
the particulate removal efficiency of the Horizontal Module increases
as a function of turndown ratio.
The Vertical TCA and PPA Modules achieve increased particulate
removal at increased flue gas pressure drop levels. For example, at
450,000 SCFM, the pressure drop across the scrubbing chamber was 14
inches of water for the TCA and 12.2 inches for the PPA. At 33% of
rated capacity with the design liquid flow rate the pressure drop
was 3.5 inches of water in the TCA and 2.0 inches in the PPA. Note
that at one-third of rated capacity with the design liquid flow rate,
the liquid to gas contacting ratios are three times higher than the
design value. However, the percentage particulate removal decreased
with decreasing flue gas pressure drop from 93% to about 70%.
In the case of the Horizontal Module, particulate removal is
primarily a function of the volume of flue gas contacted by the falling
spray droplets. As flue gas flow rate is turned down from 450,000
SCFM to 150,000 SCFM, the pressure drop across the scrubbing chamber
decreases from 1.0 inches of water to 0.10 inches of water. Since
the liquid flow rate is constant, a given amount of the gas is contacted
by three times as many droplets at one-third load as at full load.
The particulate removal efficiency was observed to increase from 92.5%
at full load to 96.5% at one-third load.
Closed Loop Water System
One of the objectives of the Test Modules Program at Mohave was
to demonstrate the operation of a closed loop water system. This means
that all the liquid water discharged from the scrubber is recycled
for use except the water which cannot be reclaimed from the waste
sludge. It should be noted that two molecules of water for each molecule
of S02 removed from the flue gas exist in the sulfate sludge as water
of hydration, or 36 pounds of water for each 32 pounds of sulfur.
Since wet scrubbers act to cool and saturate hot inlet flue gas, a
majority of the makeup water requirement in a closed loop scrubbing
system is to replace the water discharged as water vapor from the
stack.
A second objective of the Test Modules Program was to utilize
up to 75% cooling tower blowdown water in the makeup water for a
closed loop scrubbing system. The cooling tower blowdown water contained
approximately 12,000 ppm total dissolved solids. The water balances
measured during the test period are shown in Figure 10. The Horizontal
Module was capable of operating with 75% cooling tower blowdown
in the makeup and the Vertical PPA was capable of operating with 40% g
blowdown in the makeup for scale-free operation during extended testing.
Prior to converting the Vertical Module to the PPA mode, considerable
scaling occurred when the Vertical Module was operated with -65% cooling
a. The two stage PPA configuration was operated for 502 hours with 83%
cooling tower blowdown water as makeup. 340
-------
tower blowdown in the makeup water. This scaling may have been due-to
the presaturator configuration, later changed, rather than solely clue to
tha amount of cooling tower blowdown water used as make up. This prevented
obtaining satisfactory water balance data as is shown in Figure 10
for the other two scrubbers.
Both modules thus demonstrated the ability to achieve over 90% of
the makeup water discharged from the stack as pure water vapor Note
that the Horizontal Module was capable of scale-free operation at
a slurry dissolved solids level of 185,000 ppm. The Vertical PPA slurry
dissolved solids level was 70,000 ppm at the reduced consumption of
cooling tower blowdown required for scale-free operation. The. dissolved
S cn/epr!Sen^d W6re Iar9elv sodium chloride (NaCl), magnesium sulfate
(MgS04) and sodium sulfate (Na2 S04).
Table 11-2 of the March 1975 National Academy report indicated
that closed loop operation of the Horizontal Module was achieved at
Mohave, but indicated in a footnote "Mohave and Cholla experience
little rainfall and water losses due to evaporation from their sludge
ponds are significant". It is true that Mohave experiences little
rainfall, but it is not believed that a 4 gpm loss by evaporation
from the sludge pond is significant compared to the 156 gpm makeup
water requirement. The presence of rainfall had no measurable influence
on the Vertical Module where centrifuging or filtering was used as
the dewatering technique. The liquid water blowdown rate in the Vertical
Module was 6.5 gpm compared to a 160 gpm makeup water rate.
The key to closed loop operation is in the dewatering complex,
where water is reclaimed from purged scrubber .solids and returned
to the scrubbing system. This allows the makeup water flow rate to
be a function primarily of flue gas flow rate by providing all internal
water requirements with recycled water. The scrubbing system could
thus be turned down to any desired capacity while simultaneously turning
down the makeup water flow rate to avoid operating the system out
of water balance. The most critical part of the water recycle concept
required for closed loop operation was the source of pump and instrument
seal water. The seal water requirement for each module represented
a flow rate of about 50 gpm, which was independent of flue gas flow
rate, S02 content or unit load. For both Horizontal and Vertical scrubbing
systems, thickener overflow was filtered to provide the pump and instrument
seal water. In spite of the high dissolved solids in the seal water
no severe chemical scaling or corrosion was observed.
Operating Conditions at Design Gas Flow Rate
The observed operating conditions at the design gas flow rate
of 450,000 SCFM are summarized for the three types of scrubbers in
Figure 11. Several items deserve comment.
Electric Power Requirements
The flue gas pressure drop across the scrubbing chamber was an
order of magnitude less for the Horizontal Module than for the Vertical
TCA or PPA. The electric power requirements for the Vertical Module
341
-------
FIGURE 10
WATER BALANCES
INLET GAS TEMPERATURE = 270° F.
450,000 SCFM
HORIZONTAL SCRUBBER
148 6PM (eq.)
HgO VAPOR OUT STACK
PLUS THICKENER
MAKE-UP HgO-
39 GPM SERVICE HjO
117 GPM COOLING
TOWER SLOWDOWN
12,000 PPM TDS
156 GPM (25% FRESH H{0)
1
LIME SLURRY
185,000 PPM TDS
RETURN FROM POND
«•
TO POND
26 GPM
POND
4 GPM EVAPORATION
3 GPM FREE WATER IN SLUDGE
I GPM WATER OF HYORATION
8 GPM =5.1% OF MAKE-UP
VERTICAL PPA SCRUBBER
PLUS THICKENER
148 GPM (eq.)
H20 VAPOR OUT STACK
MAKE-UP H20-
96 GPM SERVICE HjO
64 GPM COOLING
TOWER SLOWDOWN
12,000 PPM TDS
•••«••••••*
160 GPM (60% FRESH HzO)
I
LIMESTONE SLURRY
70,000 PPM TDS
RETURN FROM FILTER
21.5 GPM H20
TO FILTER
33.5 GPM H20
342
6.5 GPM EVAPORATION
4.5 GPM FREE WATER IN FILTER CAKE
I GPM WATER OF HYDRATION
12 GPM =7.5% OF MAKE-UP
-------
FIGURE 11
OPERATING CONDITIONS AT DESIGN GAS FLOWRATE
450,000 SCFM
HORIZONTAL
L/6 RATIO IGPM/IOOOSCF)
SCRUBBER PRESSURE DROP (IN H20)
SYSTEM ELECTRIC POWER REQUIRED (MW)
REAGENT TYPE
SPENT SLURRY PH
REAGENT UTILIZATION
PERCENT COOLING TOWER SLOWDOWN H20
PERCENT SLURRY SOLIDS
FORMATION OF GYPSUM SCALE
(4 STAGES)
21
1.0
2.6
LIME
6.2
99%
75%
5%
NO
VERTICAL TCA
(4 STAGES)
36
14.0
3.4
LIMESTONE
5.4
75%
65%
5%
YES
VERTICAL PPA
(3 STAGES)
60
12.2
3.9
LIMESTONE
5.1
92%
41%
15%
NO
(ESPECIALLY ON FIRST
STAGE GRIDS)
-------
were consequently 30% to 50% higher than for the Horizontal Module.
For example, at design operating Conditions, the measured electric
power consumption of the Vertical Module was 3.4 MW in the TCA mode
and 3.9 MW in the PPA mode, while the power consumption of the Horizontal
Module was 2.6 MW. This included all of the power requirement including
air conditioning of the control rooms.
Reagent Comparison
Figure 11 indicates a lime utilization of 99% in the Horizontal
Scrubber. The methods used to measure this were reported in a previous
paper.2 Figure 11 also indicates a limestone utilization of 65%
with 5% slurry solids in the TCA scrubber and 92% utilization with
15% slurry solids in the PPA scrubber. The role of percent slurry
solids in influencing limestone utilization was also presented previously."
Increased slurry solids increase limestone utilization but also increase
wear on pumps and nozzles as well as increase the possibility of solids
buildup in the scrubbing system components. Both the lime and limestone
utilization data obtained in the 170 MW scrubbers compare well with
the data obtained on the pilot plant scale. Utilizing 95% pure reagents,
the data in Figure 11 can be used to calculate that 2.75 times as much
limestone by weight as lime is required to remove a given amount of SO2.
The choice between lime and limestone is a trade off between a number
of factors in which the quantity of reagent, including the associated
transportation cost, is only one factor. The operating and capital costs
of the equipment required to pulverize limestone to a 225 mesh size must
be balanced against the costs of a lime slaker to slake 1/4 inch pebble
lime. The decreased solubility of limestone, compared to lime, results
in increased holding tank sizes unless special attention is paid to the
chemistry of the scrubbing solution. Our experiments, both on the 1/2-1
MW size and the 170 MW size indicated that a higher L/G was required with
limestone than with lime to achieve the same degree of S02 removal, all
other factors being equal. Finally, the unreacted limestone increases
the amount of sludge which must be disposed of and it's presence makes
more difficult the conversion of the sludge into potentially useful
products such as gypsum wallboard. All in all, the cost per ton of a
limestone versus lime is not a true measure of the actual costs which
will be incurred.
Reheaters
It is well known that wet scrubbers cool and saturate flue gas.
Without flue gas reheat, the saturated scrubber exit gas would form
a dense, white water vapor plume. On cold days this plume can persist
for thousands of feet. Even with reheat, to 200 degrees Farenheit,
prevention of a dense water vapor plume may be impossible on some winter
mornings.
A comparison of the two methods of flue gas reheating tested at
Mohave is shown in Figure 12. On the left is the indirect reheat method,
where outside air is heated to about 375 degrees Farenheit with steam
coils and introduced into the cold flue gas at the outlet of the scrubber.
On the right is the direct reheat method, where steam coils in direct
344
-------
Figure 12
REHEATER COMPARISON FOR EQUIVALENT
REDUCTION IN FOG FORMATION
INDIRECT REHEAT
DIRECT REHEAT
550.000 ACFM
0.10 lbH20/lb Eos
(Saturated)
120,000 ACFM 'a) 75° F
0.01 lbH20/lb Air
710,000 ACFM Q) 170°F
L08 lbH20/lb Gas
*~40,000 Ib/hr Steam
Carbon Steel Coils (Finned )
Ambient Air Fan
610,000 ACFM o&190°F
0.10 IbHjO/lb Gas
550,000 ACFM 0)125° F
0.10 lbH20/lb Gas
(Saturated)
40,000 Ib/hr Steam
E-Bright
High Alloy Coils
(Smooth)
-------
contact with outlet flue gas provide the desired amount of reheat.
(The so-called direct reheat method where oil or gas is burned directly
in the exit flue gas was not tested in this program.) Both methods
of reheat require the same amount of steam for the same degree of
reduction in water vapor plume formation. Indirect reheat using relatively
dry ambient air requires an exit temperature of only 170 degrees Farenheit
to provide the same reduction in fog formation as 190 degrees Farenheit
exit temperature with direct reheat. This is because the exit gas
in the indirect reheat case is 20% diluted with relatively dry ambient
air. However, plume buoyancy or possible stack condensation problems
might dictate using a higher exit temperature gas.
During the test program, it was found that the ambient air fan
used for indirect reheat provided a built-in means of positive pressure
air purge when the scrubber was shut down. Guillotine dampers at both
ends of the scrubber were under positive pressure with ambient air
due to the indirect reheat fan, which allowed safe internal access
to the ductwork without expensive pressure sealed double guillotine
dampers. The indirect reheat steam coils were not subject to the same
environment as the direct reheat coils. These findings led to the
conclusion that indirect reheat was preferred compared with direct
reheat.
Mist Eliminators
Two types of mist eliminators were tested at Mohave, as shown in
Figure 13. The horizontal type (gas flow in horizontal plane)
utilized one mist eliminator section with front and back sides deluge
wasned at intermittent intervals. The normal wash cycle was about
15 minutes every eight hours. The vertical type utilized two mist
eliminator sections with no wash on the second section and continuous
downwash on the top side of the first section. Difficulty was experienced
in keeping this mist eliminator wash pump operating continuously during
the test program. Both mist eliminators used 3-pass blades as shown,
with plastic blades in the horizontal type and stainless steel blades
in the vertical type. The plastic blades were not capable of withstanding
temperature over about 180 degrees Farenheit without warping. The
vertical type of mist eliminator suffered the disadvantage of no capability
for removing deposits from the second mist eliminator section without
a scrubber shutdown for maintenance. Both types of mist eliminators
operated satisfactorily during the test program, but there were deposits
on the second section of the vertical-type mist eliminator blades
that continued to build up. It was concluded that a horizontal-type
mist eliminator with intermittent wash on the front and back sides
was preferred to the two section vertical-type mist eliminator with
the described wash configuration.
Unavailability - History During the Test Program
Unavailability is defined as the total time that the scrubber
was shut down and could not be operated due to design deficiency or
maintenance problems. Unavailability does not depend on whether or
not the generating unit is in operation, while availability is frequently
defined as a function of generating unit operation. In an installation
which involves a number of modules connected to one generating unit,
346
-------
Figure 13
MIST ELIMINATORS
HORIZONTAL MODULE TYPE
VERTICAL MODULE TYPE
FRONT AND BACK SIDE
WASHED AT 8 HR. INTERVALS
GAS
FROM
SCRUBBER
COLLECTED DROPLETS
FLOW BY GRAVITY TO SUMP
CONTINUOUS DOWN WASH ON.
ON TOP SIDE OF FIRST SECTION
NO WASH ON SECOND SECTION
COLLECTED
DROPLETS
FALL BY
GRAVITY BACK
TO SCRUBBER
FROM SCRUBBER
NOTE: GAS FLOW FROM SCRUBBER TENDS TO PUSH
SMALLER DROPLETS BACK INTO MIST
ELIMINATOR BLADES
-------
the unavailability record is the best indicator of how much of the
time a given module will not be available for service.
The unavailability history for the test modules program is given
in Figure 14. As can be seen, for 100% availability from a series of
modules, more modules would probably be required with the Vertical
configuration than with the Horizontal configuration.
During the Test Program, design modifications were made to alleviate
required maintenance and decrease downtime for maintenance. In the
Horizontal Module, the mist eliminator blade wash system was improved and
a single mist eliminator section was found to be satisfactory. Successful
modifications were also made to the scrubber to alleviate inlet gas
distribution problems. Methods were also successfully tested to repair
worn pump impellers and change worn spray nozzles without shutting
down the Horizontal Module. Spray nozzles made with refractory materials
were found to have longer life (estimated over one year) than nozzles
manufactured with other materials.
In the Vertical Module, many successful design modifications were
also made. However, no methods were discovered for preventing migra-
tion of the thermoplastic rubber balls between compartments without
shutting down the scrubber and redistributing the balls. Replacement
or cleaning of slurry distribution nozzles in' the Vertical Module
because of plugging or wear also required shutting down the scrubber.
Finally, scale formation at the bottom of the first stage grids was
aggravated by gas flow distribution problems in the Vertical Module.
The poor gas flow distribution at the first stage was never resolved,
but is suspected to have aggravated ball migration and gas flow channeling
problems in the Vertical Module.
DISCUSSION
There were several findings in this program which are contrary
to commonly accepted beliefs. At the beginning of the pilot plant
program it was believed by many people that it would not be possible
to reach exit concentrations of SO2 lower than 20 ppm (90% S02 removal
from an inlet concentration of 200 ppm inlet gas) because of the re-
duced partial pressure of S02 as the concentration approaches zero.
While the pilot plant results2 refuted this, the Test Modules Program
confirmed that it is possible to achieve an exit S02 concentration
of 2 ppm (99% 302 removal) on a large scale.
Many people believe today that chemistry plays an overwhelming
role in scrubber technology. Vihile chemistry is important, for example
the settling rate in the thickener tank and the quantity of sludge
produced (sulfite crystals have 1/2 molecule of water associated with
them compared to 2 molecules of water with the sulfate), one of the
test program findings was that fluid mechanics also plays an important
role, particularly in low maintenance operation. The major influences
on scrubber design and operation are segregated as shown in Figure 15.
Finally, during the Test Modules Program, twenty-four variables,
which influence visual plume opacity were identified.^ Some of these
have previously been identified (particle size, and grain loading)
348
-------
FIGURE 14
UNAVAILABILITY HISTORY-MOHAVE 170 MW TEST MODULES PROGRAM
JANUARY 16,1974 TO FEBRUARY 9,1975
HORIZONTAL HOURS
I. MODIFY AND REPAIR PLASTIC DEMISTER BLADES 503
2. CORRECT BOOSTER FAN BALANCE PROBLEMS 317
3. REPAIR CRITICAL PUMPS 256
4. REPLACE WORN-OUT SPRAY NOZZLES 238
5. MODIFY INLET GAS FLOW DISTRIBUTION 162
6. REPAIR HOPPER LEAKS 135
7. REMOVE HARDHAT FROM THICKENER 82
8. MODIFY SLAKING WATER TO PREVENT SCALE 45
9. CONDUCT INSPECTIONS FOR LONG RUNS 19
TOTAL 1757
NOVEMBER 2,1974 TO JULY 2,1975
VERTICAL HOURS
I . REPAIR GRIDS AND REDISTRIBUTE TCA BALLS 710
2. CLEAN SCALE FROM SCRUBBER INTERNALS 344
3. REPAIR OR REPLACE PLUGGED NOZZLES 153
4. REPAIR LEAKS IN TRAP-OUT TRAY 120
5. REPAIR /REALIGN PPA PACKING 85
6. CORRECT BOOSTER FAN TRIP PROBLEMS 72
7. CONDUCT INSPECTIONS FOR LONG RUNS 55
8. REMOVE HARDHAT FROM THICKENER 46
TOTAL 1585
TOTAL CALENDAR TIME 9328
TOTAL CALENDAR TIME 5813
PERCENT UNAVAILABILITY = 18.7
PERCENT UNAVAILABILITY = 27.2
-------
FIGURE 15
MAJOR INFLUENCING TECHNOLOGY
IN S02 SCRUBBING SYSTEMS
CM
Ln
o
FLUID DYNAMICS
FAN
PRESATURATOR
SCRUBBER DESIGN
SCRUBBER PUMPS
SCRUBBER PACKINGS
SPRAY NOZZLES
DEMISTER
REHEATER
DUCTING
THICKENER
VACUUM FILTER
CENTRIFUGE
PIPING AND VALVES
EROSION
INSTRUMENTATION /CONTROLS
CHEMISTRY
REAGENT FEED SYSTEM
REAGENT TYPE
PH CONTROL SYSTEM
SCRUBBER LININGS
CORROSION
SCALE PREVENTION
-------
but the experimentaloiata indicated that a major variable apparently
heretofore unrecognized was the altitude of the sun. Figure 16 indicates
the visual opacity reading which would be obtained by a "perfect"
observer if the Mohave Generating Station were located in Florida
or in Washington. With the absolute emissions of particulate matter
remaining constant, the predicted variation in opacity from 10% to
90% as shown in Figure 16 is due solely to effects of the altitude
of the sun. To date, we have not been able to convince the EPA that
the sun rises in the East and sets in the West or that the days are
longer in the summer than the winter. However, we have only been trying
£ot a year and hope for success in eliminating visual opacity regulations
in the future.
351
-------
Jx
H
M
H
55
w
w
100
90
80.
70
60
50
40
30
20
10
5:00
AM
Key West,
lorida
(June 21)
Seattle,
Washington
(Dec.21)
9:00
AM
1:00
PM
5:00
PM
LOCAL STANDARD TIME
FIGURE 16 - EFFECT OF GEOGRAPHIC LOCATION
AND TIME OF DAY ON OPACITY
352
-------
BIBLIOGRAPHY
1. Shapiro,J.L. and Kuo,W.L. "The Mohave/Navajo Pilot Facility
for Sulfur Dioxide REemoval" Second (2nd) EPA Flue Gas
Desulfurization Symposium, November 8, 1971, Newe Orleans,
Louisiana.
2. Weir,A., and Papay,L.T. "Scrubbing Experiments at the Mohave
Generating Station" 3rd EPA Flue Gas Desulfurization symposium,
May 14, 1973 - New Orleans, Louisiana.
3. Weir,A., Johnson,J.M., Jones,D.G., and Carlisle,S.T., "The
Horizontal Crossflow Scrubber", 4th EPA Flue Gas Desulfurization
Symposium, November 4, 1974, Atlanta, Georgia.
4. Handler,Philip et al "Air Quality and Stationary Source Control-
a report by the commission on Natural Resources, National Academy
of Sciences, National Academy of Engineering, and National
Research Council - Prepared for the committee on Public Works,
United States Senate - March, 1975 - Serial No. 94-4 U.S.
Government Printing Office - Catalog Number Y4,P96/10:94-4
5. Weir,A., Jones,D.G., Papay,L.T., Calvert,S. and Yung,S. "Measurement
of Particle Size and Other Factors influencing Plume Opacity"
United Nations, U.S. EPA (and other) International Conference
on Environmental Sensing and Assessment, September 14-19,1975,
Las Vegas, Nevada.
353
-------
LA CYGNE STATION UNIT NO. 1
WET SCRUBBER OPERATING EXPERIENCE
Clifford F. McDaniel, Superintendent, Air Quality Control
Kansas City Power and Light Company
P. 0. Box 211
La Cygne, Kansas 66040
ABSTRACT
This paper presents a description of the Babcock and Wilcox
designed scrubbing system and reviews the status, costs, reliability
and supportive data relating to our experiences in 1975.
355
-------
-------
LA CYGNE STATION UNIT NO. 1
WET SCRUBBER OPERATING EXPERIENCE
DESCRIPTION
The 820-megawatt La Cygne No. 1 unit began commercial operation on
June 1, 1973, as a joint project of Kansas Gas and Electric Company and
Kansas City Power & Light Company. The companies share equally in owner-
ship and output and the unit is operated by KCPL. The 630-megawatt No. 2
unit, now about 45 per cent completed, is expected to be in service by the
spring of 1977 under an identical arrangement.
The plant site is located about 55 miles south of downtown Kansas City,
one-half mile west of the Missouri state line, and was selected based on
locally available coal, water and limestone. Construction of No. 1 unit
began in 1969 and erection of the Air Quality Control System was initiated
in mid-1971.
Water for cooling purposes is furnished from a 2,600-acre reservoir
constructed adjacent to the plant site. Fly ash and spent slurry from the
AQC system is piped to a 160-acre settling pond located east of the reservoir.
Coal is delivered to the plant in off-the-road 120-ton trucks from
surface mines operated by The Pittsburg & Midway Coal Mining Co. The nearby
coal deposits are estimated to contain 70 million tons. The fuel is low
grade, sub-bituminous with an as-fired heating value of 9,000 to 9,700
Btu/lb, and an ash content of 25 per cent and sulfur content of 5 per cent.
(Exhibit A).
Limestone is obtained from nearby quarries and delivered to the plant
in off-the-road 50-ton trucks.
The boiler for No. 1 unit is a cyclone-fired, supercritical, once-
through, balanced-draft Babcock & Wilcox unit, with a rating of 6,200,000
pounds of steam per hour, 1,010 degrees F, 3,825 psig at the superheat
outlet and 1,010 degrees F at the reheater outlet. The turbine-generator
was supplied by Westinghouse and is rated at 874 MW gross output with five
357
-------
per cent overpressure and 3,500 psi throttle pressure. Three auxiliary,
oil-fired boilers are used for plant start-up or for powering a 20 megawatt
house turbine-generator. The net plant output is 820 megawatts, adjusted
to include 24 megawatts used by the AQC system and 30 megawatts by plant
auxiliaries.
PROCESS DESCRIPTION
The AQC system consists of seven identical two-stage Venturi-absorber
scrubber modules (Exhibit B) designed to treat the boiler flue gas flow of
2,760,000 ACFM. (394,300 ACFM per module at 285 degrees F). The ductwork
design does not provide for flue gas bypass of the system. Also, the plant
does not have an alternate or secondary fuel supply. Each module can be
isolated for maintenance by individual dampers. On site limestone grinding
and slurry storage facilities provide up to 1,000 tons of slurry per hour.
The unit has a balanced draft system with three 7,000 hp forced draft fans
and six 7,000 hp induced draft fans located between the AQC system and the
700 foot stack. There is a common plenum at both the scrubber inlet and
outlet. Spent slurry and fly ash are removed from the module recirculation
tank through rubber lined pipes to the settling pond at the rate of 3,500
tons of solids per day. Clear make-up water is pumped from the pond and the
loop is closed by recycling ball mill and module make-up water back into the
system.
In abbreviated terms, as the hot flue gas enters the Venturi (Exhibit C) ,
it is sprayed with slurry from 48 spray and 32 wall wash nozzles resulting
in up to 99 per cent of the particulates agglomerated to the sump below.
The gas continues through the sump making a 180 degree turn up through the
absorber section. In the reaction chamber, the SO is removed as the gas is
forced through a limestone slurry solution sprayed on stainless steel sieve
trays. The chemical reaction in part combines the calcium carbonate, water
and sulfur dioxide to form two relatively insoluble calcium salts: calcium
sulfate and calcium sulfite, which also fall to the sump. The cleaned gas
passes through demisters to remove moisture and then is reheated to avoid
deposits on the fans and provide a plume effect from the stack.
358
-------
OPERATING EXPERIENCE
As a result of the continuing modification work, the availability of
modules has been improved substantially, and for 1975, the total availa-
bility (Exhibit E) averaged 84.33 per cent, including both working and
reserve hours.
The results of a stack sampling test conducted on May 5, 1975
(Exhibit H) with the unit continuously operating from 700 to 720 megawatts
indicated an average sulfur dioxide removal efficiency of 80.14 per. cent,
and removal of 98.2 per cent of the particulates.
The ambient monitoring system indicates ground level concentrations
well below the national standards for sulfur dioxide and nitrogen
compounds (Exhibit I).
During a four hour, full load test in March (Exhibit F), the unit
maintained a minimum of 800 megawatts with all seven modules fully loaded.
The six induced draft fans were operating at less than maximum load and the
three forced draft fans had not reached maximum rating. Sulfur dioxide
removal efficiency for the six modules from which results were obtained
was 76.2 per cent.
MAINTENANCE
In addition to the modifications, the improved performance of the AQC
system has depended upon a substantial maintenance effort. Present pro-
cedures call for cleaning one module each night on a rotational schedule
and keeping all modules available during the daytime peak periods. Cleaning
requires three men from 10 to 12 hours, including time taken to open the
many sections and place holds for personal safety. Recent modifications
have greatly improved cleaning requirements and we are looking forward to
modules staying on line continuously for up to three weeks.
Areas requiring attention are reheater pluggage; demister pluggage;
Venturi wall and nozzle deposits, and sump accumulation. Hard scale has
in the past been an enormous problem, especially in the sieve trays, but by
closely controlling pH, severe scaling is not usually one of our chief troubles.
359
-------
Carry-over to the induced draft fan blades continues to be one of our
main operating concerns. We wash each individual fan about every four or
five days. This requires a fan outage of from four to ten hours depending
upon the necessity of a second or even a third washing. This is done with
very high pressure water from 4,000 to 7,000 psi. The fan is taken off
when the vibration from inbalance exceeds 12 Mils on either bearing.
To resist the effects of acid corrosion, various epoxy paints have
been used, the post recent a Plasite #4030 black paint with a durakane
resin base. This very tough surface has good erosion and corrosion
resistance but eventually cracks and should be reapplied every 8 to 10
months to maintain wheel integrity. Currently we are considering pro-
tecting those areas most affected by acid runs by cladding with Inconel 625.
MANPOWER REQUIREMENTS
The scrubber operating and maintenance force currently totals 51
people and the organization is separate from the rest of the plant (Exhibit
J). This is still considered to be a research and development situation
and once routine operation is determined, the number of scrubber personnel
will be reassessed.
COSTS
The total cost (Exhibit G) of the AQC system to date is $43 million,
or about 22 per cent of the $200 million total plant cost, or about $52 a
kilowatt installed. It is estimated that an additional $7 million investment
will be required to reach optimum system performance. These figures do not
include the allowance for funds used during construction. To date, the
City of La Cygne, Kansas, has issued $30 million in tax exempt pollution
control revenue bonds to finance the system under a lease, sublease arrange-
ment. The sale of an additional $39 million of the bonds was sold
December 1, 1975, with proceeds to be applied to the remainder of the No. 1
unit system and the AQC system for No. 2 unit.
360
-------
La Cygne production costs for 1975 for energy averaged 8.82 mills
per kilowatt hour. The limestone and operating and maintenance expense of the
AQC system totaled 1.26 mills, or about 14 per cent of the production
costs. Labor accounted for .45 mills, limestone .55 mills, and other
costs were .27 mills.
The total annual cost of producing power including fixed costs from
La Cygne No. 1 unit in 1978 is estimated at 15.24 mills per kilowatt hour.
Interestingly, the comparative cost for the No. 2 unit, which will burn
Wyoming coal supplied by Amax Coal Company from mines in the Gillette
area, will be 21.96 mills per kilowatt hour. As these projections indi-
cate, total cost of producing power from No. 1 unit, even with the
complicated air quality control system, will still be 44 per cent less
expensive than for No. 2 unit. Of the total cost, the air quality component
for La Cygne No. 1 unit is 3.03 mills per kilowatt hour, and for No. 2 unit
is 1.77 mills per kilowatt hour. Unit No. 2 will be equipped with an
electrostatic precipitator.
ADDITIONAL MODIFICATIONS
Planned additional modifications include the installation of an
eighth module to be in service the summer of 1977. This is estimated to
cost $5.2 million and will improve the cruising capability of the unit
from 700 megawatts to 800 megawatts. The original AQC system design
allowed space for this modification. Changes in the induced draft fans are
also being made to reduce or eliminate a fan paralleling problem. Many
other experiments and minor changes are continuing to improve performance
and to reduce operating and maintenance expenses.
361
-------
LA CYGNE STATION
COAL AND ASH ANALYSIS
COAL
Proximate
Volatile
Fixed Carbon
Ash
Moisture
100.00
BTU/lb.
9421
Grindability 59.59
Ultimate
Moisture
Carbon
Hydrogen
Nitrogen
Chlorine
Sulfur
Ash
Oxygen
8.60
51.93
3.43
0.94
0.027
5.39
24.36
5.33
100.007
ASH
Analysis
Phosphorous Pentoxide 0.15
Silica 46.05
Ferric Oxide 19.23
Alumina 14.07
Lime 6.86
Magnesia 1.02
Sulfur Trioxide 7.85
Potassium Oxide 2.48
Sodium Oxide 0.60
Titania 1.02
Other 0.67
100.00
Fusion Temperature
Reducing I. D. 1957
Soft (H=W) 2045
Soft (H=W/2) 2169
Fluid 2321
Oxidizing I. D. 2156
Soft (H=W) 2338
Soft(H=W/2) 2415
Fluid 2520
Exhibit A
362
-------
X
zc
CO
La Cygne limestone wet scrubbing system
-------
FIGURE I - LACYGNE1 FGD MODULE
REHEAT
STEAM
550 °F
(-
1
REHEAT
1000 PPM -
-*- SO2 -j
190° F H
_C_OILS H
f n n-n o o o
x
-43"
HsO
J*-i ' '
OT Ai R
— FROM
BOILER
INTERMITTENT
OVERSPRAY
2150 GPM
CONTINUOUS
UNE5ERSPRAY
140 GPM
VENTURI SPRAY
/\ /\ /\ /\
WALL
WASH
SPRAY
v V \/ V V
VENTURI
THROAT
FLUSH
PREDEMISTER
ABSORBER
SPRAY
200-6OO GPM
RECIRCULATION
CA.CO3 70 G/L
CASO3 35 G/L
CASO4 25 G/L
FLYASH 4O G/L
P H - 5 5 - 6 0
8- 10% SOLIDS
T O
FAN
SPENT SLURRY
TO POND
700 GPM
«35OO TONS/ DAY
93000 TONS/ YEAR
53 ACRE FEET/YEAR
LIME
STONE
SLURRY
FEED
20%
SOLIDS
VENTURI
RECIRC PUMP
5000 GPM
TOTAL FOR ALL MODULES
EXHIBIT C
ABSORBER
RECIRC. PUMP
90OO GPM
364
-------
LA CYGNE SCRUBBER WATER ANALYSIS
CATIONS
CALCIUM HARDNESS SOL. (Ca)
MAGNESIUM HARDNESS SOL. (Mg)
SODIUM (Na)
POTASSIUM (K)
COOLING
_LAKE
123.2
11.0
12.0
4.5
SETTLING
POND
696.0
48.0
22.0
23.0
ANIONS
BICARBONATE ALK (AS HCO )
CHLORIDE (Cl)
SULFATE (SO )
SULFITE (SO )
SILICA (Si02)
109.8
24.6
247.9
* ND
2.4
36.6
177.8
1627.3
* ND
20.6
OTHERS
pH (pH UNITS)
CONDUCTIVITY IN MICROMHOS
7.87
649.0
7.0
4380.0
*ND - Not Detected
Exhibit D
365
-------
MODULE AVAILABILITY SUMMARY
LA CYGNE 1975
TOTAL
GENERATION
BOILER CAPACITY
MONTH
January
February
March
April
May
June
July
August
September
October
Novemb er
December
A B
C
D E
F
G AVAILABILITY*
MWH HOURS FACTORS
Turbine Generator Repair
Turbine Generator Repair
82.4 96.03
Generator
94.6 85.1
87.8 85.4
78.4 89.7
74.64 88.07
78.43 83.62
66.16 77.26
Generator
92.87 90.79
Generator
90.72 87.39
89.5
Repair
94.2
83.9
89.6
87.29
84.38
46.27
Repair
80.18
Repair
80.87
76
.6 92.96
91.5
96
89.33
25 Days
89
84
83
78
84
73
.5 89.8
.9 84.1
.7 85.4
.01 92.44
.67 78.72
.62 71.91
89.3
86.1
87.4
85.00
77.71
73.07
83.4
88.6
85.2
83.06
74.24
64.69
89.4
85.8
85.6
84.07
80.25
67.57
7
244
23
332
324
297
294
239
74
,886
,873
,014
,526
,952
,870
,402
,954
,660
694
683
667
590
630
610
231
38.8
52.7
53.2
47.2
46.7
39.3
24.3
15 Days
93
.18 96.09
89.39
93.94
90.83
165
,058
346
50.6
17 Days
85
.20 86.89
88.56
83.67
86.19
278
,597
597
46.8
*Working Hours -4- Reserve
Hours In Month
Exhibit E
-------
Four Hour Load Test Results
(Test for Maximum Credited Megawatts)
Date: March 24, 1975
Time: 6:10 pm to 10:25 pm
Power: Min. - 800 MW (e) Gross - Peak* - 830 MM (e)
Outside Temp: 26 F
Data
F
Gas Flows (cfm) 380
Throat Pos.
Reheat Temp. ( F)
Absorber Pump
Venturi Pump Flow 4
Venturi AP
Reheater AP (X2)
Abs-Dera. AP
Hot Air Damper Pos.
(% Open)
Reheat Outlet Damp.
Pos. (% Open)
Scrubber Outlet Press.
(-"H2o)
I . D . F an Amps
I.D. Fan Inlet Damper
Pos . (% Open)
F.D. Fans Amps
Lab pH**
Conductivity**
Sulfite (g/1)
Carbonate (g/1)
S02 Efficiency
Inlet (ppm)***
Outlet (ppm)
,000
22"
170
ON
,000
8.0
4.0
-
46
51
360,000
22"
180
ON
4,000
7.5
3.5
5.5
23
37
3 8. 3 (common)
540 520
100
540
5.95
2100
41.1
94.4
75.1
5700
1419
98
500
5.89
2200
36.2
84.1
79.1
5138
1075
400,000
22"
180
ON
5,000
-
3.0
-
44
98
620
98
490
5.80
2000
45.9
69.4
72.2
5516
1533
410,000+
22"
150
ON
4,000
11.0
3.5
8.5
-
84
34.0
600
100
5.68
2250
35.4
84.4
72.0
4995
1220
400,000
22"
150
ON
5,000
10.0
1.0
8.5
48
100
380
100
5.92
2000
45.1
91.9
7
5700
1
390,000
22"
170
ON
5,000
11.0
4.5
9.0
32
96
380
23
5.79
2100
29.4
109.0
82.9
5017
857
340 ,000
22"
185
ON
5,000
9.0
12.5
9.0
64
100
5.73
2200
54.7
76.6
75.7
5120
1243
Exhibit F
367
-------
COSTS
LA CYGNE STATION
Scrubber Operating Expense 1975
OPERATING LABOR
OPERATING MATERIALS
MAINTENANCE LABOR
MAINTENANCE MATERIALS
LIMESTONE
TOTAL
$ 601,029
195,926
416,206
386,397
1,256,048
2,855,606
0.265 Mils/KWH
.086
0.184
0.171
0.554
1.260 Mils/KWH
TOTAL UNIT
SCRUBBER PORTION
CAPITAL COSTS
$200,000,000
43,000,000
$244/KW
$ 52/KW
COAL: $7.10 TON - 37.7
-------
Module Evaluation
Test f *
DATE:
TIME:
LOAD RANGE:
AMBIENT TEMP:
IA CYGNE STATION
STACK SAMPLING TEST
5/15/75
11:30 P.M. to 8:00 P.M.
700 to 720 MW Continuous
72° F
(Data taken by independent
test group)
AVERAGE SO2 REMOVAL: 8^- ,
PARTICULATS REMOVALS: 98.2% Inlet 9.9# per 10 BTU
BAROMETTRIC P.: 29-2* Outlet .18* per 10 BTU
BEL. HUM. DATA LOGGED AT
gas Flow Indicated
Throat Position
Reheat Tenraeroture
Absorber Pumo
Venturi Slurry Flow
Vent or i ^ P
Reheater & p
Absorber Eerr.. A ?
Hot Air Daarper Pos.
^ Ct>en
Reheat Outlet Daaper
Pos. 4, Open
Scrubber Outlet Fress.
C-"HpO)
I.D. Fan Ar.ps (Control
RuC'IIi )
I.D. Fan Ar.cs (Breaxer?
I.D. Fan Inlet Dampers
Pos. <& Orcen
F.D. Fan Aaps (Control
Room)
F.D. Fan Arans (Breaker)
Lab T5H/C.R.
Sulfite (,-z/l)
Carbonate (p/1)
S02 Efficiency %
Inlet (DUB)
Outlet (-or*-)
A
320K
22
170
On
4,000
7
4
-
46
40
38
460
60
5.95
53,9
155.0
76
4,506
1,068
B
280 K
22
190
On
4,000
6
3
4
20
35
480
60
6.02
37.0
85.3
81
4,297
834
C
300 K
22
170
On
4.000
7
3
-
50
55
460
60
5.03
37.4
108
81
4,663
892
D
340 K
77
148
On
4,000
7
8
8
_
75
28
400
60
5.80
58.4
75.3
82
4.273
776
E
320 K
77
180
On
4,000
8
6
S
50
82
410
60
5.98
61.6
109
77
4.982
1.121
F
300 K
97
160
On
4,000
7
11
f>.
94
100
440
32
5.90
54.7
93.1
83
4 1 S6
704
G
300 K
. 22 ..
iqo
On
4.000
8
8.
8
100
100
5.90 i
64.8
95.9
81
L Sli
Q1 7
BOILER DATA:
F. D. Fan Discharge ("
Air Flow %
Fuel Flow %
Feedwater Flow #/Hr.
Excess Op %
Windbox-Furn. Diff. Press.
Furn. Pressure
Sec. Super Gas Pressure
Pend. Reheat Gas Pressure
Primary Super Gas Pressure
44
68
72
5,200,000
2.5
31
-2.0
-3.4
-3.2
-6.0
Horz. Reheat Gas Pressure
Econ. Outlet Gas Pressure
Feedwater Pressure
Throttle Pressure
Throttle Temperature
Hot Reheat Temperature
Air to Air Heater A
B
Air from Air Heater A
B
Gas to Air Heater A
B
Gas from Air Heater A
B
Exhibit H 369
-7.3
-10.0
4,100
3,500
970
1,000°F
155.5
lib.9
561.0
598.0
655.5
668.7
308.7
30S.9
-------
LA CYGNE STATION
AMBIENT MONITORING SYSTEM
STATION 1
STATION 2
STATION 3
MILES FROM PLANT
PRIOR TO START UP
CONCENTRATION S02 - ppm
10
12.5
RECENT LEVELS
CONCENTRATION S02 - ppm
N02 - ppm
NATIONAL STANDARDS
CONCENTRATION SO- - ppm
NO- - ppm
.009
.019
.030
(80mg/M3)
.008
.009
.028
.024
.030
.050
(100mg/M3)
.003
.029
.030
LEAD PEROXIDE CANDLES
21 CANDLES RING THE STATION AT DISTANCES FROM 2 to 6 MILES. 3 MORE CANDLES
ARE LOCATED ABOUT 30 MILES NORTH OF THE PLANT, JUST SOUTH OF THE KANSAS
CITY METROPOLITAN AREA.
SULFATION GRADUALLY INCREASED FROM .029 TO .031 mg/CM2 PRIOR TO
2
COMMERCIAL PLANT OPERATIONS AND HAS SINCE INCREASED TO .082 mg/CM .
Exhibit I
370
-------
LA CYGNE AIR QUALITY CONTROL
MANPOWER REQUIREMENTS
OPERATORS PER SHIFT
3 Attendants 13
3 Clean-Up 14
1 Shift Foreman 5
1 Process Attendant (Chemist) 1
Exhibit J
371
33
MAINTENANCE
Mechanics 8
Apprentice Mechanics 2
Welder 1
Electrician 1
Technician 1
Plant Helpers 2
Foreman 1
16
ADMINISTRATIVE
Superintendent 1
Engineer ]_
2
TOTAL 51
-------
RECENT SCRUBBER EXPERIENCE AT THE LAWRENCE ENERGY CENTER
THE KANSAS POWER AND LIGHT COMPANY
D. M. Miller, Manager, Electricity Production
Kansas Power and Light Company
818 Kansas Avenue
Topeka, Kansas 66612
373
-------
Recent Scrubber Experience at the Lawrence Energy Center
The Kansas Power and Light Company
I would like to share with you our participation in the practical application
of the Combustion Engineering Air Pollution Control Systems installed on two steam
generating units located at The Kansas Power and Light Company Lawrence Energy Center.
These steam generators were both supplied by Combustion Engineering and are capable
of firing pulverized coal, fuel oil, or natural gas, each or all together.
My Company is The Kansas Power and Light Company headquartered in Topeka,
Kansas. We operate wholly in the State of Kansas. Our peak load this past summer
was just short of 1% million KW. Our principal electric generating stations are
located at Tecumseh, Hutchinson and Lawrence, Kansas. Over the past 25 years we
have primarily burned natural gas supplemented by fuel oil and/or coal when the gas
supply has been interrupted. These interruptions have gone from 8-12% in the 50's
and 60's to over 50% in 1975 and we would expect these interruptions co exceed 65%
this year. This shift from gas to fuel oil and coal is not a shock and is an occu-
rance we had planned for - it is just here a little earlier than we had programmed.
Even though in 1967, 65% of our steam generators were equipped to burn pul-
verized coal, we were still classified as a gas fired utility. In 1967 when the
decision was made to build the 400 MW addition at our Lawrence Energy Center, it
was quite clear that natural gas would not be available to fulfill its fuel require-
ment. A 20 year contract for Kansas coal with delivery starting in 1967 was nego-
tiated to fulfill the primary fuel requirement for this unit. Natural gas and fuel
oil would supplement the coal burning the first years of operation after start up
in the spring of 1971.
374
-------
Also in 1967 when the decision to build this coal burning addition was made,
it was assumed that by 1971 there would be some ambient and/or emission regulations
In effect for particulate matter and sulfur dioxide. Based on this assumption and
the availability of 3-4% sulfur, 12% ash coal, the decision was made to install as
original equipment, facilities to treat the stack gas to remove both fly ash and
sulfur dioxide. This was accomplished by the installation of a Combustion Engineering
Limestone Injection-Wet Scrubber System.
Since the startup of the 400 MW unit was scheduled for 1971, it was further
decided to retro-fit a similar system on an existing 125 MW unit at Lawrence so
that we could gain some operating experience before starting up the larger unit.
This smaller unit went into service in the fall of 1968 and has provided the in-
dustry with many innovations which are now being used on newer installations being
planned in the United States.
The Combustion Engineering Air Pollution Control System at Lawrence is designed
to remove both fly ash and sulfur dioxide simultaneously from the flue gas stream.
Both the scrubbing of the fly ash and the chemical reaction to absorb the S02 are
accomplished in a common vessel.
I would like to describe for you the system as it was initially installed on
our 125,000 KW unit. (Figure 1)
The limestone injection equipment is the existing pulverizers where the rock
and coal are pulverized together and are blown into the furnace as a mixture. The
•et scrubber system contains the spray nozzles, the marble bed, the demisters and
reheater coils. The heat source for the reheater is extraction steam from the
turbine through the existing feed water heater stream. New I.D. fans were installed
to handle the increase in head requirements.
The actual system has two one-half capacity scrubbers each with its own inlet
duct and damper, bypass duct and existing stack.
375
-------
The bypass system was installed so that we could keep the steam generator
in service on fuels other than coal so that we could maintain reliability of KW
output while learning how to operate the Air Pollution Control System. We have
been able to do just that as we have operated many hours with scrubbers out of
service. This has allowed us much opportunity to maintain and modify this system
while generating KWH which still is our primary job.
The dirty pot overflow water from both scrubbers containing about 2% solids
is pumped to the settling pond where the solids are separated. The clarified
water is returned through an overflow structure to the clear side of the pond and
to the scrubber supply pump intake structure. These pumps supply water to the
scrubber system and the cycle repeats. The supply pumps and ponds will be common
to the 125 and 400 MW units.
The scrubber on Unit 4 was started up in 1968 with the configuration shown
in Figure 2. This was a simple system but presented many operating problems and
shortcomings:
Problems encountered were:
1. Build up and plugging of the inlet duct where hot gases
entered the scrubber.
2. Erosion of the scrubber walls.
3. Corrosion of the scrubber internals
4. Plugging and scaling of drain lines, tanks and pumps
5. Plugging and scaling of the marble bed
6. Plugging of spray nozzles
7. Plugging of demister due to carryover from bed
8. Plugging of reheater due to carryover through demister
9. Buildup on ID Fan rotors due to moisture getting through
reheater under abnormal conditions resulting in fan
unbalance.
376
-------
In addition to the operating problems the S02 removal was quite low due to
the everburning of the limestone and the dropout of the lime with the ash in the
bottom of the scrubber. Satisfactory particulate removal was achieved however.
After the first winter's operation, the scrubber was revised to the configu-
ration shown in Figure 3. The addition of soot blowers in the inlet duct and re-
heater helped the plugging problems there. The demister was raised to increase
the distances between the bed and demister. The pot overflows were directed to
the storage pond and a large recycle tank and pump were installed to accumulate
and pump the highly alkaline bed underflow back into the bed. New type spray
nozzles were installed to minimize plugging. The bottom part of the scrubber tanks
were lined with gunite to reduce erosion and corrosion of the walls. We borrowed
the coal pulverizers from an adjacent coal fired unit and pulverized the limestone
in one of these mills and blew the additive material into the furnace on through
the observation doors at the arch level of the boiler. This did provide a reactive
•aterial for use in the scrubbers.
Most of the problems were reduced but not eliminated by the first set of
revisions. The recirculation system did improve the S02 removal. Maximum S02
removal was possible under this mode of operation but the resulting sulfur scale
formations in the bed drain lines and pumps would soon put the scrubber out of
business.
Revisions were made to the scrubbers the following summer in an attempt to
further minimize the corrosion, erosion, scaling and plugging problems. Figure 4
shows the configuration of the scrubber in October, 1970. Major revisions made
(ere:
1. Sandblasting and coating the interior of the scrubber vessels with
a two coat glass flake lining.
2. Replacement of all internal steel pipe systems with plastic and
fiberglass.
3. Replacement of stainless steel demisters with fiberglass demisters.
377
-------
4. Addition of a ladder vane under the bed to improve gas distribution.
5. Modification of the pot overflow drain piping to allow the drains to
return to the recycle tank allowing for a semi-closed loop operation.
6. The original copper fintube reheat coils were removed and replaced
with a carbon steel fintube coil. The copper units plugged easily
and the fins were flattened by the soot blower jets.
7. . We replaced the 14" drain outlets from each scrubber with a 2' x 4'
sluice way to alleviate the drain pluggage we had been experiencing.
All the above modifications helped in reducing operating problems except for
the demister plugging. Scaling could be controlled by maintaining close ph control
but only at the expense of reduced SO removal. Manual washing of the demisters
was necessary on an every other night basis to maintain load capability of the
unit.
We operated the scrubber in this manner with a fair amount of success through
the winter of 1970-71 and into the spring of 1971. We were able to by frequently
in$pecting the system and rather continually performing small maintenance jobs,
keep the scrubber system available to meet the required coal burning during this
period.
We were not satisfied with the reliability of the system at this time but
planned to start up the 400 MW Lawrence addition under this mode of operation.
I will discuss that operation later in this report.
We continued our fight to alleviate the nagging pains of clogging the recycle
nozzles or the recycle piping system and more carryover from the bed than the
demister could reasonably handle which plagued us with a dirty reheater which did
not lend itself to simple repeated cleaning.
Then came the operation of the winter of 1971-72 when we were also burning
coal on the 400 MW Lawrence addition.
The only thing we could consistently do was make sulfur bearing scale —
on the scrubber walls — in the scrubber marble bed and on the demisters of all
the scrubbers in both units.
378
-------
Operation of the scrubber systems remained in this mode until the late winter
of 1972. At that time Combustion Engineering completed tests in their laboratory
scrubber at Windsor using a high solid slurry crystallization process to control
saturation and precipitation to eliminate scale within the scrubber. The system
was quite successful in the lab so a crash program was started to modify the
Lawrence 4 scrubber for full scale testing of the high solids system.
The temporary revamp took about 30 days to complete and the scrubber was
tested for 60 days during May-June, 1972 using the high solids slurry system to
control scaling. The results were encouraging; so encouraging in fact that the
decision was made to spend the summer of 1972 to rebuild and modify the scrubber
systems on both Lawrence 4 and 5 to operate with the high solid-crystallization
concept.
The modification included the installation of:
1. Large crystallization tank
2. 4' enlargement of the scrubber bed on #4
3. New rubber lined pipe slurry system
4. New plastic spray nozzles
5. New bed drain system
6. New slurry pumps and strainers
7. New multiple mixers in tank
8. New two bank fiberglass demister with a power wash system
This mode of operation is shown on Figure 5.
The operation of this system on Unit 4 through 1973 was much improved over
anything we had been through in the past. We still experienced some of the nagging
maintenance and cleaning problems such as:
1. Isolated corrosion
2. Unsatisfactory damper operation
3. Expansion joint failure
379
-------
4. Demister fouling
5. Rapid slurry pump wear
6. Valve failures
but were able by the end of 1973 to maintain the scrubber system operable with
2-8 hour shifts of manual cleaning per scrubber per week.
This cleaning requirement lessened a little into 1974 when for economical
reasons in the cost of coal, we completed negotiations for a supply of low sulfur
coal from SE Wyoming for these units. We started receiving this coal: 10,000
BTU/#, .5% sulfur and 11% ash in the fall of 1974 and had phased out all the SE
Kansas coal by late spring 1975. The bypass duct system in the Lawrence 4 unit
was removed at this time due to complete deterioration.
The operation of the scrubber system on the Wyoming coal has proved to be
somewhat better and more economical due to a lesser amount of sulfur removal re-
quired. The scrubber system is still operating in the high solids mode as an S02
and particulate removal system. Our normal manual cleaning requirements have been
reduced to 2-4 hour shifts per scrubber per week.
In 1974 this unit was available for operation 343 days. 50% of the fuel
consumed was coal, 2% fuel oil and 48% natural gas.
In 1975 this unit was available for operation 333 days. 64% of the fuel
consumed was coal, 3% fuel oil and 33% natural gas.
I would estimated this unit to be available for operation 330 days in 1976
with the following fuel split; 80% coal, 8% fuel oil and 12% gas.
Remember the scrubber system must be in operation to burn coal.
You may have heard that we are replacing this scrubber system with a rod
section followed by a spray tower and wonder why. You have a right to that
question.
380
-------
We are presently replacing this successful scrubber system because since
1968 in modifying and revising the scrubber modules and operating many hours at
corrosion levels that were bad, we actually consumed the physical scrubber plant.
Now that we have achieved - at the Lawrence Energy Center and under the conditions
of operation there - a mode of operation that we can maintain, we have had so
much deterioration of vessel and equipment that we must build a new system to
have one.
We are extremely interested in the new system since parts of this system
will be installed at our new Jeffrey Energy Center which will be discussed later.
So much for the Lawrence #4 Unit, lets now review the operations of the
tawrence #5, 400,000 KW unit - installed in 1971.
We followed the programs as discussed in Lawrence 4 with the scrubber system
on this unit. We started with recycle to the bed - we made scale, we modified
and installed the high solids mode of operation, we added 2 scrubber modules and
we operated burning coal when it was required.
We experienced the problems of Lawrence 4 on this unit with one additional
problem, a more severe distribution of flue gas to and through the marble beds
on #5 than on #4. This has complicated our lives on #5 unit and has been a large
part of the reason our operating successes on #5 have come slower and not reached
as high a plane as #4.
However we achieved the following performance with Unit #5.
In 1974 - the unit was available for operations 338 days - consuming 27%
coal, 6% fuel oil and 66% gas.
In 1975 - the unit was available for operations 352 days - consuming 42%
ieoal, 13% fuel oil and 45% gas.
I expect this unit to be available for service 330 days in 1976 - consuming
Wl coal, 25% fuel oil and 15% gas.
381
-------
We are studying the future of the scrubber system on Unit 5 and my best
guess today is that it will follow in the footsteps of Unit 4 and will be converted
to the rod and spray tower system.
These experiences at our Lawrence Energy Center are the basis for our con-
clusions concerning the stack gas clean-up system we intend to install at the
Jeffrey Energy Center. (Figure 6)
The Jeffrey Energy Center is composed of 4 - 680,000 KW coal fired units
planned to be on line 1 unit early 1978, 1 in 1980, 1 in 1982 and 1 in 1984.
All of these units will be fired by coal now under contract from the AMAX Coal
Company from the Powder River Basin in Wyoming. The coal specifications for this
plant are designed to be 8000 BTU/#, .32% sulfur and 5.8% ash.
We have purchased from Combustion Engineering a stack gas clean-up system
that will remove in excess of 99% of the particulate matter, and 60% of the S02
in the flue gas. An overfire air system at the tangential fired pulverized burners
will control tk» BMI emission.
The stack. fM&ft&Mn-** •*»«•* (Fig«r« 7) is composed of an electrostatic
praeipitator following the air heater them the induced draft fans and then through
spray towers. Th« effluent gas fro« the spray towers will be mixed with some hot
gas from the ID fan discharge for reheating the flue gas as it flows to and out the
600 ft. chimney.
We are of the opinion that we will be able to operate this system under the
conditions we will have at the Jeffrey Energy Center in such a manner to be good
neighbors to those who live around us in Pottawatomie County, Kansas and to satisfy
the existing requirements of The Department of Health and Environment of the State
Of Kansas and the Environmental Protection Agency.
382
-------
ENGINEERING
AIR POLLUTION CONTROL SYSTEM
MNSIS ROWRUICKT COMPANY
L«ren« Station Unit No 5
OU*
STACK
CAS
SCRUBBER
-------
•
. ! H AIR
1
.
1
LAWRENCE N* 4
DECEMBER 1966
SCRUBBER
SOLID DISPOSAL '
Oi
CO
•u
WATER
///////////////
OEMISTER
SPRAY
TO STACK
N* 4
APCS
')8ER 1970
SCRUBBERS (2)
CLARIFIED
POND
DRAIN TANK
• SOOT BLOWER AIR
WATER WASH LANCE
**•
LAWRENCE N* 4
CE-APCS
OCTOBER 1972
SCRUBBERS (2)
(ENLARGED DEPTH-4')
RECYCLE TANK (I)
C ENI. AROEO)
CHAIN TANK
CLARIFIED
FROM
PONO
TO
SOLID
DISPOSAL
r(J) POND
-------
-
o
' •
---|!M!!J!|a|!!!|l" "'"'.^^'^" """T
™ ** 1 mt, MJHI - M. OB.™ 1
-------
INTRODUCTION TO: DOUBLE ALKALI
FLUE GAS DESULFURIZATION TECHNOLOGY
Norman: Kaplan
Industrial Environmental. Re&earch Laboratory
Office of Research and Development
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ABSTRACT
The chemistry and significant process and design factors applicable
to sodium/calcium double alkali systems are presented. Technical
terminology associated with these systems is defined.
Estimates of capital and operating costs are presented in addition
to reasonable targets for reagent consumption in double alkali systems.
The EPA development program is summarized and a status of technology
is presented.
387
-------
NOTES
1. Company Names and Products.
The mention of company names or products is not to be considered
an endorsement or recommendation for use by the U.S. Environmental
Protection Agency.
2. Consistency of Information.
The information presented was obtained from a variety of sources
(sometimes by telephone conversation) including system vendors,
users, EPA trip reports and other technical reports. As such,
consistency of information on a particular system and consistency
of information between the several systems discussed may be lacking.
The information presented is basically that which was voluntarily
submitted by developers and users with some interpretation by
the author. The order of presentation of information or the amount
of information presented for any one system should not be construed
to favor or disfavor any particular system.
3. Units of Measure.
EPA policy is to express all measurements in Agency documents
in metric units. When implementing this practice will result
in undue cost or difficulty in clarity, IERL/RTP is providing
conversion factors for the particular non-metric units used in
the document. Generally, this paper uses British units of
measure.
For conversion to the Metric system, use the following equivalents:
British Metric
5/9 (°F-32) °C
1 ft 0.3048 meter
1 ft2 0.0929 meters2
1 ft3 0.0283 meters^
1 grain 0.0648 gram
1 in. 2.54 centimeters
1 in.2 6.452 centimeters2
1 in. ^ 16.39 centimeters^
1 Ib (avoir.) 0.4536 kilogram
1 ton (long) 1.0160 metric tons
1 ton (short) 0.9072 metric tons
1 gal. 3.7853 liters
388
-------
INTRODUCTION AND BACKGROUND
At the last EPA Symposium on Flue Gas Desulfurization (FGD),
held in Atlanta in November 1974, plans were announced for an
EPA co-funded full-scale utility boiler double alkali (D/A)
demonstration program in which the system vendor and utility
would be selected on a competitive basis. At that time there
were two industrial boiler applications and an industrial kiln
control application of the technology in this country. Japan
was a bit more advanced in the application of this technology
in that there'was one 150 Mw application of the technology in a
full-scale utility system. As a result of the increased testing
of D/A systems in this country in various pilot plants, at a
20 Mw utility prototype system at Gulf Power, at various industrial
boiler systems and other industrial applications (sulfuric acid
plants, kilns,etc.) and the interest generated by EPA's request
for proposal (RFP) for a full-scale utility demonstration of D/A
technology, this technology now appears to be competitive with
lime/limestone wet scrubbing for some utility applications.
In response to the RFP, EPA received three viable proposals
for installation of double alkali systems at full-scale coal-
fired utility systems, ranging in size from 192 to 575 Mw.
In Japan there are now three operational full-scale utility
applications ranging in size from 150 to 450 Mw.18 Additionally,
Kureha has developed another variation of the D/A process which,
reportedly, is a significant improvement over their limestone D/A
system that they now have operating as a full-scale utility system.
This variation is called the "Sodium Acetate-Gypsum" process and
uses sodium acetate as the S02 absorbent, limestone as the regen-
erant, and produces a gypsum by-product. This process is currently
being pilot tested at a-5000 Nm3/hr (1 Mw) pilot plant using a
perforated tray tower operating at a 280 mm (11 inches) H20 pressure
drop, L/G of 7-8 gal./lOOO ft3 gas with an S02 inlet concentration of
1400 ppm and an exit concentration of less than 10 ppm. This
process will be discussed in greater detail in a later paper in
this session.
"Double alkali',' or "dual alkali" processes as they have come
to be known, like their precursors the lime or limestone wet scrubbing
processes, are aqueous alkali scrubbing processes used for FGD,
If a "black box" view of these processes is employed, they are,
with minor exception, like the lime/limestone scrubbing systems:
389
-------
lime or limestone is consumed, and a calcium sulfite/sulfate and
fly ash wet solid waste product is produced. A more detailed
examination inside the "black box," however, would show that
the overall process has been split into a number of intermediate
steps designed to improve upon the lime/limestone processes by
increasing reliability of operation, utilization of lime and/or
limestone, and sulfur oxide removal efficiency and, under certain
circumstances,'producing a solid waste with better handling
characteristics. Whereas in a lime/limestone process the absorption
of the SO from the flue gas and the production of the waste product
occur to some extent simultaneously -in a single reaction system,
in the double alkali processes, these two steps are separated through
the use of an intermediate soluble alkali; absorption and production
of waste product can then occur in separate system components.
Separating the absorption and waste production functions
accomplishes two very important objectives First, it permits
scrubbing the flue gas with a soluble alkali thus limiting the SOX
absorption reaction only to gas/liquid chemical equilibrium and to
the rate of transfer of SOX from the flue gas to the scrubbing
solution. In lime/limestone processes the rate of dissolution
of lime or limestone is a third important limiting factor. Thus
SO absorption efficiency in a double alkali system is potentially
higher than in a lime/limestone system with the same physical dimensions
and liquid/gas flow rates. Additionally, soluble calcium is minimized
and calcium slurry is kept out of the scrubbing apparatus thus
preventing solids scaling and plugging in this critical area. Another
benefit of this is better control of the lime/limestone reaction with
acidic sulfur containing compounds in separate equipment specifically
designed for this reaction thus potentially increasing lime/limestone
utilization.
Although a number of other processes can technically be considered
double alkali processes, this paper is limited to consideration of the
sodium/calcium based double alkali processes. In these processes, a
soluble sodium based alkali (NaOH, Na2S03, Na2C03, NaHC03) is used to
absorb SOX from the flue gas in the scrubber, and then a calcium based
alkali (Ca(OH)2, CaO, CaO^) is reacted with the SOx-rich scrubber
effluent liquid to precipitate the insoluble CaS03-l/2 H20, CaSO^-2H20
and mixed crystal and regenerate sodium based soluble alkali for recycle
to the scrubber system. Double alkali systems with an ammonium/calcium
base have been tested: while they might have advantages in the reaction
with calcium compounds, their main disadvantage is the potential for
pollution by a visible ammonium salt plume from the scrubbing apparatus
caused by the highly volatile ammonium compounds. Another variation
that might be considered double alkali is Monsanto's "Calsox" process,
which uses an aqueous organic base as the absorbent solution in combination
with lime as the calcium-supplying alkali to produce the throwaway product.
390
-------
DliFIN'TION AND DISCUSSION OF TERMS
As with any specialized technology, a discussion of flue gas
desulfurization in general, and double alkali technology more
specifically, involves the use of special terminology which" has
evolved with the technology. While terms are understandable to those
dealing with the subject on a daily basis, they can be somewhat
ambiguous to others. To clarify some of these ambiguities, and
to define terms used here and by others describing double alkali
technology development, a number of terms and concepts are defined
and d Lscussed in a general sense.
Absorption/Regeneration Chemistry
The main chemical reactions that take place in double alkali
systems can be divided functionally into the absorption and regeneration
reactions. A number of secondary reactions which have very important
effects on the overall functioning of the system also take place.
These include oxidation, softening and sulfate removal reactions which
are discussed under the appropriate headings.
The regeneration reactions and in some cases the absorption
reactions will be dependent upon which calcium supplying regenerant
is used—lime or limestone. With lime the system can be operated
over a wider pH range than witli limestone. This wider pH range allows
lime systems to operate over the complete range of active alkali
hydroxide/sulfiLe/bisulfite whereas limestone systems can only
operate in the suIfite/bisulfite range.
The main overall absorption reactions are described by the
following oc|unl ions : .
2NaOll + SO., > Na.,S03 + t^O
Na,,SO + SO t- II 0 •> 2NaHSO^
2. j t- £• ^
The main overall regeneration reactions are described by the following
equations for lime and limestone respectively:
Lime^
Ca(0ll)2 + 2NaHS03 > Na2S03 + CaSCyl/2 H20* + 3/2 H20 (3)
Ca(OII)., + Na2S03 + 1/2 H^O -> 2NaOH + CaSO^l/2 H^ (4)
Ca(0nf + Na SO + 2H 0 J CaSO, -2H?0 * + 2NaOH (5)
\_* rt V W 1 I / ' \ ' !»<-*.v./vy, n -^ ^ ^
391
-------
LJ mo stone
CaCO + 2NaHS0
1/2
3 -I- CaScyl/2
+
(6)
D/A systems are frequently and erroneously referred to as
sodium ion scrubbing systems. It should be stated that, from a
"pure" chemistry viewpoint, the reactions presented in equations
(1-11) and (15-18) do not involve the sodium ion (Na+) ; however,
the presentation is made using compounds of sodium because sodium
systems are prevalent in D/A applications and because this allows
showing the reactions using electrically neutral reactants and
products rather than charged ions. For example, the absorption
reactions involve reaction of S02 with an aqueous base such as OH ,
SOo, HCOo or 663 rather than with Na+ which does not take
part in the reaction, but is only present to maintain electrical
neutrality. Thus equations (1) and (2) for example could have been
written as equations (la) and (2a) respectively:
2 OH" + S02 ->
$0
H20
H20
21IS0
(la)
(2a)
Active Alkali
This term is the sum of the concentrations of NaOH, Na2C03?
Na.,50 and NaHSO^ in the scrubbing solutions. Sodium bisulfite is
included in this definition although it is not technically an alkali
(i.e., it cannot react with SO in.these systems); however, it can be
converted to an alkali by reaction with lime or limestone. It should
also be noted that the molar capacity of each of these species for
absorption of SO,, is different, and can vary from zero to 2 moles of
S0? per mole of active alknli. This difference in molar capacity for
absorption of SO., is illustrated by the following reaction equations:
Na-CO., + 2SO- + H_0 -•• 2NaHSO + CO
4. -J ~ *- -* *-
(sodium carbonate molar capacity: 2 moles SO /mole)
NaHC0 + S0 -
(sodium bicarbonate molar capacity: 1 mole SO /mole)
NaOH + SO " NaHSO
(sodium hydroxide molar capacity: 1 mole SO /mole)
Na S03 + SO 4- HO -> 2NaHSO
(sodium sulfite molar capacity: 1 mole S02/mole)
NaHSO- 4 SO ->- No reaction
(sodium bisulfite molar capacity: zero mole S02/mole)
392
(7)
(8)
(9)
(10)
(ID
-------
Molar capacity is simply the number of moles of S0? needed to convert
1 mole of the absorbent alkali completely to sodium bisulfite. Since
there is a difference in the molar capacity of different active alkali
components to absorb S0? , active alkali is a descriptive rather than a
quantitative term. If the concentration of each of the active alkali
components (moles/liter) is known, however, the capacity of the scrubbing
liquor to absorb SC>2 (moles of 80,,/liter of solution) can be calculated
as the sum of each of the active alkali component concentrations
multiplied by their respective molar capacities as follows:
Scrubber liquor SOo capacity (moles/liter) =
2 [Na2C03J + iNaOHj + [NaHCO-j] +
TOS
This is an abbreviation for "total oxidizable sulfur." It
denotes the concentration of sulfur compounds in solution in which
the sulfur is in the -t-4 oxidation state. Simply, this is the total
concentration of sulfite plus bisulfite.
TOS (moles/liter) = [SO =] + [HSO ""]
Sulfate is not part of TOS, since the sulfur is in the +6 oxidation
state in this specie. Sulfur dioxide dissolving in scrubbing solutions
increases l'ie TOS in solution.
Active Sodium
This is the concentration of sodium in solution which is
associated witli the active alkali.
[Na+] = [NaOH] +'2 [Na.,COj + [NallCO.. J + [NaHSO ] + 2 [Na SO ]
act / j J -3 ^ J
If NaOH, NatlCO , Na CO , NaHSO or Na S03 solids are added to double
alkali solutions, the increase in sodium ion in solution is "active
sodium." If Na SO or NaCl, for example, is added, the increase in
sodium ion is "inactive sodium," Active sodium is not increased by
the dissolution uf S0? in scrubber solutions. Note that the term
"active sodium" can be misleading in that the sodium ion doesn't
participate in any of the process reactions.
393
-------
Oxidation
Oxidation in a double alkali system refers to the conversion of
TOS to sulfate by one of the following equations:
HSO ~ + 1/2 0 ->• S0,= + H* (12)
S03= + 1/2 02 > S04= (13)
Simple oxidation of S0? to SO. in the flue gas is also considered
oxidation in the double alkali system:
SO, + 1/2 02 * S03 (14)
Oxidation in the system has the effect of changing active sodium
to inactive sodium, or active alkali to inactive alkali.
Oxidation may occur in any part of the system: in the scrubber,
the reaction vessels, or in the solids separation equipment. In
general, rate of oxidation in the system is thought to be a function
of rate of dissolution of oxygen, pH of the scrubbing solution,
impurities present in solution and concentration of reactants. Oxida-
tion rate is thus affected by composition of the scrubbing liquor
(scrubbing liquors containing high concentrations of dissolved salts
may absorb oxygen more slowly), oxygen content of the flue gas,
impurities in the coal and lime or limestone, and the design of the
equipment, (the regeneration and solids separation sections of the
system in particular can be designed to limit dissolution of oxygen
and number of scrubber contact stages is extremely important).
Oxidation rate is expressed as a percentage and is calculated from
an overall material balance on .the system:
Oxidation rate (%) = [SO ~ leaving the system (moles)] ^QQ
[Total sulfur collected (moles)]
Sulfate leaving the system is total moles of sulfate in the solid waste
plus any sulfate in the associated liquor.
Sulfate Regeneration
This term is a misnomer. What is really meant is sulfate removal
from the system with regeneration of active alkali from inactive sodium
sulfate. (See Sulfate Removal.)
394
-------
Sulfate Removal
Sulfate is removed from the system with regeneration of active alkali
from inactive sodium sulfate. Examples of these sulfate removal reactions
are given below:
Na.SO. 4- Ca(OH)9 + 2H 0
/ 4 i 2.
2NaOH + CaSO -2H004-
42
(gypsum)
(15)
Na.-SO.
/ 4
2CaSO -1/2' H00 + H0SO. + 3H_0 ->• 2NaHSO, + 2CaSO, -2H-OI (16)
3 i 2 4 2. 3 4 /
(gypsum)
y Na?SO. + x NaHSO + (x+y) Ca(OH). + (z-x) HO >
£f ^T J £ *~
(x+2y) NaOH + x CaSO -y CaSO, 'z tt^Q^
(mixed crystal or solid solution)
(17)
Na0S0/
24
3H00
2
ElGCtr°lytic
> 2NaOH
1/2 0
(18)
Sulfate removal should be accomplished in an environmentally acceptable
manner; a simple purge of soluble Na2SO/ from the system to land or water-
way disposal is not acceptable in large ^quantities in most cases.
Softening
This term is used to describe various methods used to lower the
dissolved calcium ion concentration in regeneration solutions. The
purpose of softening the scrubbing liquor before recycling to the
scrubber is to assure that it is subsaturated with respect to gypsum.
This reduces the gypsum scaling potential in the scrubber. Following
are examples of softening reactions:
+ Na2C03 -> 2Na +
+ Na2S03 + 1/2 ti^ -»• 2Na+ + CaSCyl/2
(19)
(20)
(21)
In each of the above reactions, calcium ions are removed from solution
as part of an insoluble material, outside the scrubber system. Reactions
(19) and (21) are referred to as carbonate softening. Reaction (20) is
considered sulfite softening. Generally, dilute systems employ carbonate
softening while concentrated systems resist scaling due to the high
sulfite concentrations which prevent high Ca++ ion in the scrubber liquor.
395
-------
Dilute vs. Concentrated System
Dilute or concentrated refers to the active alkali concentration
in a particular system. This differentiation is made because, in
theory at least, based on their solubility products in water, both
CaSO, and CaSO, should not precipitate from a solution of sulfite
and sulfate simultaneously when using relatively small quantities
of lime slurry for regeneration, unless the concentrations of sulfite
and sulfate are present in a certain ratio. This can be shown by
dividing one solubility product equation by the other:
tea"""] [S04=] = KsPl
[Ca J [S03 ] = Ksp2
I *
from this, cancelling Ca ion concentration,
[so4=j
-— = constant
iso3~]
The ratio of sulfate to sulfite for simultaneous precipitation of
CaSO. and CaSOo is shown to be a constant.
The constant in the above equation is the ratio of solubility
product constants of calcium sulfate and sulfite. Then, in theory, if
the ratio of sulfate to sulfite is higher than this constant, only
calcium sulfate should precipitate; and if the ratio is lower than the
constant, only calcium sulfite should precipitate.
This very simplified consideration of the chemistry given above
is clouded in the "real world" by factors that contribute to non-
ideal behavior in these systems. These factors include changes in
ionic activities in solutions containing high electrolyte concentrations
and evidence of coprecipitation of calcium sulfite and sulfate in the
form of a "mixed crystal" or "solid solution" in a manner which is not
completely understood at present. >3
With due consideration to the non-ideal behavior of these systems,
however, under given conditions, a ratio of sulfate to sulfite in
solution can be determined at which the previously cited examples hold.
The ratio establishes a definition for "dilute" or "concentrated"
double alkali systems. When the ratio is such that either gypsum or
both gypsum and calcium sulfite will precipitate from the solution with
the addition of slaked lime, then the system is "dilute."
396
-------
Lime or Limestone Stoichiometry
Lime or limestone stoicbiometry can be expressed as a percentage,
based on an overall material balance around the system:
Lime Stoichiometry =
Limestone Stoichiometry =
moles CaO added
mole sulfur collected
moles CaCOn added
100
mole sulfur collected
100
Lime or Limestone Stoichiometry is an indication of the efficiency of
usnge of lime or limestone in the system. Ignoring alkali components
in the flyash collected and the alkalinity added with sodium make-up
to the system, 100% Stoichiometry is complete utilization' stoichiometriee
over 100% represent less efficient utilization of lime or limestone.
Stoichiometries under 100% indicate alkalinity from ocher sources.
Feed S t o i c h iome t ry
Tins is calculated hy a material balance around the scrubber.
ll is usually expressed as the ratio:
ir , , i q, • .1 = .".j--i"-"r--SQo Capacity (moles/liter) x Flow (liters/min)
r e eel o L o i (-11 <- /-»/^ / i i • \ ——^—^——^-~—
SO,, (mole/min)
Tliis ratio is evaluated for streams
entering the scrubber.
Feed Stoichiometry is a measure of the ability of the incoming
liquor to react with or absorb all of the incoming SO in the
scrubber, assuming ideal contact of gas and liquor. Feed Stoichiometry
above 1.0 is required for high SO,, removal. At feed Stoichiometries
at or below 1.0, assuming ideal contact between the gas and liquor,
there will be significant equilibrium S07 partial pressure above the
liquor, and thus S0? removal is theoretically limited to the value
calculated on the basis of this SO,., partial pressure in the exiting
flue gas.
397
-------
SIGNIFICANT PROCESS AND DESIGN FACTORS
A commercial double alkali system must be designed to remove the
desired quantity of sulfur oxides from a given flue gas stream, while
operating in a reliable manner and discharging environmentally acceptable
solid waste product. In fulfilling these design objectives, cost is
also an important factor.
SO? Removal
The fact that small quantities of sulfur dioxide can be removed
from large amounts of relatively inert gas by cyclic processes involving
absorption into aqueous solutions of sodium sulfite/bisulfite has been
known for some time. Johnstone et al. published a paper in 1938 giving
data on the vapor pressure of S02 over solutions of sulfite/bisulfite
and methods of calculating these equilibrium values under various
conditions. The equilibrium partial pressure of S02 above sulfite/bisulfite
solutions, the theoretical limit which a practical design can approach,
is generally a function of solution temperature, pH, concentration of
sulfite/bisulfite and total ionic strength. Since Johnstone's work a
number of organizations have pursued this technology with laboratory,
pilot plant and full scale applications for flue gas desulfurization,
and many have demonstrated the ability for high removal efficiencies.
(It should be noted that although Johnstone's work was aimed at cyclic
processes with thermal regeneration, such as the Davy Powergas system,
the vapor pressure data is also applicable to double alkali systems
which use chemical regeneration.)
Once methods have been established to determine equilibrium SO
vapor pressure over scrubbing solutions, of the various concentrations
to be encountered in an operating system, it becomes a matter of
standard chemical engineering practice to design adequate gas absorption
equipment to accomplish the desired SO- removal in a system. For
comparison, it should be noted that the design of lime/limestone slurry
absorption equipment is further complicated by the kinetics of dissolution
of the lime or limestone, the particle size of the suspended material,
and the crystal morphology of the lime or limestone.
Reliable Operation
System reliability can be adversely affected by two classes of
problems, chemical and mechanical.
The mechanical problems include malfunction of instrumentation
and mechanical and electrical equipment such as pumps, filters,
centrifuges, and valves. These problems in a commercial FGD system
can be minimized by careful selection of materials of construction and
equipment and by providing spares for certain equipment such as pumps
398
-------
and motors which are expected to be in continuous operation and are
prone to failure after a relatively short period of operation.
Another important consideration in minimizing mechanical problems
is the institution of a good preventive maintenance program.
The chemical (or physical/chemical) problems which may be
associated with a double alkali system include scaling, excessive
sulfate build-up, production of poor-settling solid waste product,
water balance and build-up of non-sulfur solubles which enter the
system as impurities in the coal or lime. Each of these factors
is associated with reliable system operation, or production of
an environmentally acceptable solid waste.
a- Scaling - One of the primary reasons, and probably the most
important, for development of double alkali processes was to
circumvent the scaling problems associated with lime/limestone
wet scrubbing systems. Therefore, a double alkali system should be
designed to operate in a non-scaling manner.
Scaling is caused by precipitation of calcium compounds from
process liquors, on the surfaces.of various components of the
system. When this occurs in the scrubber it is particularly
troublesome since the flue gas path through the scrubber, if
affected, could cause shutdown of the boiler-scrubber system and
lower reliability.
Since scrubbing in double alkali systems employs a clear
solution rather than a slurry, there is a tendency to ignore
potential scaling problems. Testing experience with double
alkali systems has indicated, however, that scaling can occur
and indeed the problem should be a legitimate concern in the
design of any system. Bo.th gypsum and carbonate scale build-up
has been recognized in these systems. Gypsum scaling is caused
by the reaction of soluble calcium ion with sulfate ion formed
in the system through oxidation of the absorbed sulfur dioxide
or from absorbed sulfur trioxide according to the reaction:
Ca^ + S04= + 2H20 -> CaS04'2H20 + (22)
In dilute systems gypsum scaling is controlled by softening the
regenerated liquor prior to recycling to the scrubber while in con-
centrated systems gypsum is not a problem since the high sulfite concen-
tration keeps the Ca++ ion low. Softening ensures that the liquor
recycled to the scrubber system is unsaturated with respect to gypsum;
therefore, with proper softening even if some sulfate is formed xn
the scrubber, the liquor will not be saturated with gypsum and cause
399
-------
scaling on the inside surfaces of the scrubber. In concentrated active
alkali systems, a special softening step is not necessary since high
sulfite concentration is maintained throughout the system. This sulfite
maintains a low calcium ion concentration (sulfite softening), and thus
maintains the scrubbing solution unsaturated with respect to gypsum.
Based on experience gained in lime/limestone scrubbing testing,
a certain factor of safety in the prevention of gypsum scaling probably
exists in double alkali systems. Gypsum has been found to
supersaturate easily to about ,130% saturation; thus, even if sulfate
formation is higher than expected, gypsum may not precipitate in
the scrubber until the liquor is over 130% of saturation with respect
to gypsum.
Carbonate scaling usually occurs as a result of localized high
pH scrubbing liquor in the scrubber where C0£ can be absorbed from
the flue gas to produce carbonate ion. This ion subsequently
reacts with dissolved calcium to precipitate calcium carbonate
scale according to the following series of reactions:
Carbon dioxide absorption by high pH liquor:
C02 4- 2 OH~ - C03= + H20 (23)
Calcium carbonate scaling:
C03= + Ca++ •* CaC03 + (24)
Based on experience with the General Motors full scale double
alkali system, 5 carbonate scaling could occur with scrubber liquor
pH above 9. At lower pH, the carbonate/bicarbonate equilibrium
system tends to limit the free carbonate ion and thus prevent
precipitation of calcium carbonate:
,H+ + C03= C HC03" (25)
Thus, carbonate scaling can be eliminated by control of pH in the
scrubber.
b. Production of Poor-Settling Solids - Under certain conditions,
the waste product solids produced in the regeneration sections of
various double alkali systems have a tendency not to settle from the
scrubber liquors. This creates problems in the operation of settlers,
clarifiers, reactor clarifiers, filters and centrifuges. Although
400
-------
the phenomenon has been observed in the laboratory testing conducted
by EPA on dilute systems and in the laboratory and pilot plant work
conducted by Arthur D. Little, Inc. (ADL) on dilute and concentrated
systems, it is not completely understood, but is thought to be a
function of reactor kinetics.^
Some of the factors thus far identified which appear to affect
the solids settling properties are reactor configuration, concen-
tration of soluble sulfate, concentration of soluble magnesium and
iron in the liquor, concentration of suspended solids in the reaction
zones, and use of lime vs. limestone for reaction. Based on laboratory
work in dilute systems (about 0.1 M active sodium) using limestone,
it appears that solids settling characteristics degraded significantly
at soluble sulfate levels above 0.5 M. Based on laboratory work
with concentrated systems (about 0.45 M TOS, 5.4 pH, 0.6 M sulfate)
using limestone, marked degradation of solids settling properties
occurred at a magnesium level of 120 ppm and virtually no settling
of solids occurred at the 2000 ppm magnesium level. Equal degradation
of solids settling properties also occurred in concentrated systems
when the sulfate level was raised to the 1.0 M level while maintaining
low magnesium level (about 20 ppm) and keeping other variables constant.
Envirotech6 advocates the recycle of precipitated solids from
the thickener underflow to the reaction zones in an effort to grow
crystals which settle faster and are more easily filtered.
ADL cites reactor configuration as being important in the ^
production of solids with good settling and filtration characteristics.
Their basis for this is comparative tests of a simple continuous flow
stirred tank reactor (CFSTR) with the ADL/Combustion Equipment
Associates (ADL/CEA) designed reactor system under similar conditions.
The ADL/CEA reactor system appeared to give better settling solids
over a greater range of conditions than a simple CFSTR.
c Reliability/Availability Data - Three full-scale utility
and several industrial boiler double alkali systems have started
operation in Japan since 1973. Although detailed availability data
is not shown here for these units, few, it any, significant operating
problems have been reported for them. Ando has given more detail
on these systems in another paper.-1
401
-------
In the U.S. the 20 Mw utility prototype system constructed by
CEA./ADL for Southern Services at Gulf Power's Scholz station has
been in operation for almost a year. That system started up in
February 1975 and operated intermittently through January 1976. It
was snut down for modification and repairs from mid-July through
mid-September 1975. Excluding that 2-month period, the system
maintained a 90% availability during the first, approximately 1 year
of operation. Most of the problems experienced were related to
the filtration system.
FMCrs double alkali kiln control system has been in operation
at their Modesto, California, plant for over 5 years. That system
is equivalent to 10 Mw based on gas rate or about 30 Mw based on
the size of the regeneration system and S02 absorbed. As of approx-
imately a year ago, that system reportedly was available for control
of SOX emissions from the barium reduction kilns 100% of the time
required. The kilns were in service 95% of the time with scheduled
shutdowns for routine maintenance and/or change of chemical process
service every 3 or 4 months.
Environmentally Acceptable Solid Waste
Double alkali systems should be designed to produce an environ-
mentally acceptable end product. Some of the properties which can be
ascribed to such a solid waste product are listed below:
-Non toxic
•Low soluble solids content, non-leachate
•Low moisture content
•Non-thixotropic
-High compressive or bearing strength
402
-------
In order to generate waste product solids which have properties
approaching those listed, the following design considerations are
appropriate:
•Sulfate removal
•Water balance/waste product (cake) wash
•Gypsum vs. calcium sulfite
•Sludge fixation technology
a. Sulfate Removal - In double alkali systems, some of the
sulfur removed from the flue gas takes the form of soluble sodium
sulfate due to oxidation in the system, thus changing some of the
active sodium to the inactive variety. When sodium in the system is
converted to the inactive form (Na^O^) , it is relatively difficult
to convert back to active sodium. To convert inactive sodium to
active sodium, sulfate ion must be removed from the system in some
manner, while leaving the sodium in solution. The alternative to
this is to remove the sodium sulfate from the system at the rate
it is being formed in the system. This alternative is not desirable
since it is wasteful of sodium and generally is carried out by
allowing the sodium sulfate to be purged from the system in the
liquor which is occluded in the wet solid waste product.'' The solid
waste product can then potentially contribute to water pollution due
to leachability. Water run-off can lead to contamination of surface
water, while leaching and percolation of the leachate into the soil
can result in contamination of the ground water in the vicinity of
the disposal site. Failure to allow for sulfate removal from
double alkali systems will ultimately result in a) precipitation
of sodium sulfate somewhere in the system if active sodium is made
up to the system, or b) in the absence of make-up, eventual deterioration
of the SOo removal capability due to the loss of active sodium
from the system.
Equations (15), (16), (17) and (18), shown previouslv under the
definition of "sulfate removal" describe several sulfate removal
techniques which have been used in FGD system pilot tests.
The first equation depicts the sulfate removal technique
used in dilute active alkali systems:
Na2S04 + Ca(OH)2 + 2H20 J 2NaOH + CaS04'2H2(H (15)
(gypsum)
Concerning the full-scale dilute alkali system installed and
operating at the Parma, Ohio, transmission plant of General Motors,
and dilute systems in general, Phillips^ stated:
403
-------
"The presence of Na2SO^ in the scrubber effluent
is the prime factor influencing the design of the
regeneration system. Na-^SO^ is not easily regenerable
into NaOH using lime. The reason being that the
product, gypsum, is relatively soluble. . . . ^280^
cannot be causticized in the presence of appreciable
amounts of S0-j= or OH~ because Ca"1"4" levels are held
below the CaSO^ solubility product. To provide for
sulfate causticization, the system must be operated
at dilute 01I~ concentrations below 0.14 molar. At
the same time, S0^= levels must be maintained in the
system at sufficient levels to effect gypsum
precipitation. ... We selected 0.1 molar OH~ and
0.5 molar S0^= as a design criteria."
In a previous paper, Phillips^ showed a plot of equilibrium caustic
formation in Ca(OH)2~Na2S04 solutions at 120°F which is the basis
for selection of the design criteria. The essence of this discussion
is that if the active sodium concentration is sufficiently dilute,
sulfate can be removed from the system by simple precipitation as
gypsum by reaction of lime with sodium sulfate.
Since, as explained above, this reaction will not proceed .to a
great extent in concentrated active alkali systems, other techniques
must be employed to effect sulfate removal in these, systems.
The second equation depicts a technique which is used in the full
scale double alkali systems in Japan, and which has been pilot tested
by ADL under contract with EPA:
Na2S04 + 2CaS03-l/2 H20 + H2S04 + 3H20 •> 2NaHS03 + 2CaS04'2H20-t- (16)
(gypsum)
This technique is used to precipitate gypsum by dissolving calcium
sulfite in acidic solution thus increasing the Ca"^" in solution enough
to exceed the solubility product of gypsum. Ideally according to
equation (16) 2 moles of gypsum should be precipitated for each mole
of sulfurie acid added. In practice, however, this is not the case
since any material which functions as a base can consume sulfurie
acid and reduce the efficiency of this reaction for its intended
purpose.-^ Unreacted lime or limestone, sulfite ion and even sulfate
ion can consume sulfurie acid thus lowering sulfate removal from the
system.
404
-------
Conceivably, this method of sulfate removal may be economically
unattractive in applications with very high oxidation rates, and
where the gypsum produced must be discarded. The economic picture is
considerably changed where this system is used merely as a slip-
stream treatment to supplement other sulfate removal methods and/or
where the solid product gypsum is saleable as is the case in Japan.
The third equation describes a phenomenon which has been
referred to as mixed crystal or solid solution formation:
x NaHS03 + y Na2S04 + (x+y) Ca(OH)2 + (z-x) H20 ->• (17)
(x+2y) Na01l'+ x CaSO-j-y CaS04-z H20 4-
(mixed crystal or solid solution)
This phenomenon is described by R.H. Borgwardt of EPA2 as it applies
to lime/limestone wet scrubbing based on pilot plant investigations.
A similar phenomenon has been observed by ADL in some of their early
pilot testing of double alkali systems in conjunction with CEA, and
later in the EPA/ADL dual alkali test program.
It appears that under certain conditions the solids precipitated
in lime/limestone and double alkali systems contain sulfate, sulfite
and calcium; however, the liquor from which these solids precipitated,
appears to be subsaturated with respect to gypsum. This is based on
the fact that pure gypsum crystal could be dissolved in the mother
liquor from which the mixed crystal/solid solution was precipitated.
In addition, the solid material was examined by X-ray diffraction
and found to contain no gypsum; infrared analysis confirmed the
presence of sulfate.
Borgwardt found that the molar ratio of sulfate to sulfite in
these- solids was primarily a direct function of sulfate ion activity
in the mother liquor. In pilot test work with lime/limestone
scrubbing, with little or no chlorides present and normal magnesium
level (below 1000 ppm) in solution, the sulfate to sulfite molar
ratio in the mixed crystal solids was found to reach a maximum level
of 0.23. This is equivalent to a [S04=]/total [SOX] ratio in the
solids of '0.19.
In pilot test work with concentrated double alkali systems, ADL
observed the simultaneous precipitation of sulfate and sulfite with
calcium in lime and limestone treatment of concentrated double alkali
scrubbing liquors. This phenomenon was surprising at first, in light
of the reasoning which led to the development of dilute double alkali
405
-------
systems; i.e., gypsun cannot be precipitated from solutions
containing high active alkali concentrations. It was a simple
technique for sulfate removal in concentrated systems. The [SO^j/
total [SO,.} ratio observed in pilot double alkali work was as
..
high as 0.20.-5 Coincidentally, this is the same value
observed by Borgwardt in lime/limes tone testing. This leads
to the suspicion that the same phenomenon is occurring in both
processes. The mother liquor from which these solids were
precipitated was also found to be subsaturated in gypsum, and when
the solids were examined, pure gypsum was not found.
Based on the observed data, it. appears reasonable to design
a concentrated active alkali system for a particular situation in
which the system oxidation rate is below about 20%. In this case,
sulfatc can be removed at the desired rate, without the necessity
for purging Na2S04 or supplementing the system with other complex
methods of sulfate: removal.
The fourth equation shows sulfatre removal as sulfuric acid
in an electrolytic cell:
Electrolytic
Na2S04 + 3H20 cell ^ 2Na01I + H2S04 + H2 + 1/2 02 (18)
This method is the basis for operation of the Stone and Webster/
Ionics process sulfate removal technique. In Japan, Kureha/Kawasaki
has pilot tested the Yuasa/Ionics electrolytic process for sulfate
removal in conjunction with their double alkali process. They feel
that this process will be less expensive overall than the presently
;;scd sulfuric acid addition method. In addition, they feel that
sodium losses from the system can be cut in half through the use of
this methodt from 0.018 moles Na loss/mole S02 absorbed to 0.009
moles Na loss/mole S02 absorbed.
Another approach to sulfate control is to limit oxidation. With
sufficient: limitation of oxidation, by process and equipment design,
it may be possible to control sulfate by a small unavoidable purge
of Na2S04 with the solid waste product. To design for minimum oxidation,
there should be minimum residence times in equipment where the
scrubber liquor is in contact with oxygen-containing flue gas, and
all reactors, mixers, and solids separation equipment should be
designed to minimize absorption of oxygen from air. In addition,
it has been reported^ that oxidation of scrubber liquors can be
minimized by maintaining very high ionic strength. One possible
406
-------
explanation for this is that high ionic strength liquors are poor
oxygen absorbers and that oxidation in these systems is oxygen
absorption rate limited.
b. Water Balance and Waste Product (Cake) Washing - In order
to operate a closed system to avoid potential water pollution problems,
system water balance is a primary concern. Water cannot be added to
the system at a rate greater than the normal water losses from the
system.
Generally fresh water is added to a D/A system to serve many
purposes. These include:
•Saturation of flue gas
•Pump seal needs
'Demister washing needs
•Slurry make-up needs
•Waste product washing
•Tank evaporation
On the other hand water should only leave the system in the
following ways:
'Evaporation by the hot flue gas
•Water occluded with solid waste product
•Water of crystallization in solid waste product
Careful water management, part of which is the use of recycled
rather than fresh water wherever possible, is necessary in order to
operate a closed system.
As previously indicated, disposal of wet solid waste containing
soluble salts is ecologically undesirable. In addition, allowing
active alkali or sodium salts to escape from the system is an
operating cost factor. Sodium make-up to double alkali systems is
usually accomplished by adding soda ash (recently quoted at $70 per
ton) at some point in the system. Thus, both ecological and economic
considerations dictate that waste product washing is desirable.
A rotary drum filter, belt filter or centrifuge is usually the
equipment in which the final solids separation is made. This
equipment can be designed to permit solids washing with fresh water.
407
-------
One concern in waste product washing is the extent to which the
cake should be washed. At present, there are no federal regulations
concerning the amount of total dissolved solids (IDS) which is permitted
in a waste product which is to be disposed of as landfill. The future,
however, may unfold stringent regulations in this area. One obvious
consideration in waste product washing is system water balance. Un-
limited waste product washing is not possible if a closed system operation
with no liquid stream discharge is a goal. Another more subtle reason
for limiting waste product washing is the potential problem of non-sulfur/
calcium solubles build-up in the system. These non-sulfur/calcium
solubles enter the system with the fly ash, flue gas, the lime and/or
limestone and the make-up water. Of these, probably the soluble
material in highest concentration would be sodium chloride which results
from the absorption of I1C1 from the flue gas by the scrubber solution.
A material balance around the system at steady state necessitates
that solubles leave the system at the rate they enter. Thus, depending
upon how well the waste product is washed, a certain level of non-sulfur
solubles will be established in the system. Since the only mechanism
for these solids to leave the system is as part of the wet solid
waste, a certain purge rate is necessary. This purge also necessitates
the loss of some sodium from the system. Practical limitations in
filter design and water balance probably would limit a system to two
or three "displacement washes" of the waste product (one displacement
wash means washing with an amount of fresh water equivalent to the
amount of water contained in the final wet waste product per unit of
waste). Depending on the characteristics of the waste product and
the design of the washing system, one displacement wash can reduce
the solubles content of the waste product by as much as 80%.
c. Gypsum vs. Calcium Sulfite - Although gypsum (CaS04'2H20)
is more soluble than calcium sulfite hemihydrate (CaS03'l/2 H20),
gypsum may be considered a more environmentally acceptable end
product. The solubility of gypsum in water is about 0.25%; that
of calcium sulfite is on the order of 0.0025%. It is interesting to
note that while gypsum is a naturally occurring mineral, calcium
sulfite is not found in nature. In the solid waste product (sludge)
from lime/limestone and double alkali FGD systems, gypsum has better
handling properties than calcium sulfite. Sludges containing a
high ratio of gypsum to calcium sulfite are less thixotropic, better
settling, more easily filtered, and can be more completely dewatered
than sludges containing a high proportion of calcium sulfite. Another
important characteristic which has been attributed to high gypsum (as
opposed to calcium sulfite) sludges is their higher compressive or
bearing strength. Typically, lime/limestone systems which generate
408
-------
solids having a high proportion of calcium sulfite can be filtered
to 40-50£ solids, while some double alkali systems producing high
gypsum cake can be filtered to over 657, solids.10
Some explanation for the behavior of these sludges is given by
Selmeczi and Knight.11 Although filter cakes appear dry, they still
contain a considerable amount of water and thus, upon vibration or
application of stress, they have the tendency of again becoming fluid.
This thixotropic property and high moisture content are both explained
by the morphology of calcium sulfite clusters. Due to the highly
open, porous or sponge-like nature of these clusters, a considerable
amount of water is retained in the clusters. The calcium sulfite
crystals are rather fragile and break under pressure releasing some
of the water, which results in the sludge becoming fluid.
It is possible in some double alkali systems to produce a
gypsum product. In Japan, where by-product gypsum is a-saleable
product, the calcium sulfite solids produced are oxidized completely
to gypsum in a separate oxidation process tacked on to the tail end
of the system. In applications where high excess combustion air is
present in the boiler or where low sulfur coal is burned or a
combination of these conditions is present, the oxidation rate in the
system tends to be high (possibly about 90%) and the proportion of
gypsum in the sludge tends to be high. In some dilute systems the
proportion of gypsum in the sludge can be increased by augmental
aeration of the scrubbing liquor.1^ Crystal seeding techniques used
in conjunction with augmental aeration can produce relatively coarse
grained gypsum crystals with good'-dewatering and structural properties
in the final waste product.
d. Sludge Fixation Technology - Chemical or physical fixation of
the sludge produced in a double alkali system is another potentially
important means of producing an environmentally acceptable solid waste
product. This technology is under investigation by T.U. Conversion
Systems, Inc., Chicago Fly Ash, Dravo Corporation, and Chemfix
Corporation. Most of their efforts are concentrated on sludge produced
from the more prevalent lime/limestone systems; however, there is also
some evaluation of double alkali sludges. The objective of sludge
fixation technology is the production of a non-toxic, unleachable
solid waste product which has reasonably high load bearing strength.
If double alkali sludges are amenable to this type of treatment, the
need to reduce soluble sulfates in the solid waste product becomes
409
-------
less of a problem. Some sodium sulfate has been found to be
physically or chemically tied up In the solid calcium sulfate/
sulfite crystal lattice;^ however, the extent of this phenomenon is
not generally considered to be adequate to remove all of the sodium
sulfate produced via oxidation. Sludge fixation technology may
be a mechanism by which additional sodium sulfate can be removed
from the system without adverse environmental effects. There is
some concern as to whether this is viable, however, since sludge
fixation chemistry involves pozzolanic reactions between calcium
compounds and flyash components in the sludge which may only
involve multivalent ;ions rather than monovalent sodium. In other
words, monovalent ions such as sodium may either a) not take part
in the pozzolanic reactions, or b) inhibit or limit such reactions.
Further investigation ,is called for in this area.
Dilute vs. Concentrated System
The selection of a dilute or concentrated double alkaJi system
is an important design consideration in any application. In general
it can be stated that the concentrated systems are more suited
to applications in which oxidation is expected to be relatively
low, and that, conversely, dilute systems are favored in applications
where oxidation rates are high. High sulfur Eastern coal applications
on utility boilers where excess air is controlled carefully and
maintained at the lowest value consistent with complete combustion,
the concentrated systems may be favored. On the other hand, in
utility or industrial boiler applications where Western low sulfur
coal is burned, and/or where control of oxidation is difficult due
to high excess air, the dilute systems may be favored.
Oxidation rate is promoted when low sulfur coal is burned, since
the ratio of oxygen to sulfur dioxide in the flue gas is higher than
in high sulfur coal applications. Since oxidation is a strong function
of the rate of absorption of oxygen, liquor which is dilute in TOS
Is subject to having a greater proportion of these species oxidized
by a given amount of absorbed oxygen than one in which the TOS is
more concentrated.
Under a given set of conditons without consideration given to
waste disposal, a concentrated system can be installed at lower
capital cost than a dilute system as previously discussed; however,
the desirability to produce a manageable solid waste (dilute systems
can be designed to produce high gypsum sludges) could, In cases,
override the capital cost issue.
410
-------
Sodium Consumption
Sodium consumption is an important performance criterion for
a double alkali system not so much from the viewpoint of economics
as from the viewpoint of producing a leachable waste product as
stressed under sulfate removal. Sodium consumption is only a
minor factor in the operating cost of a system representing about
2% of the annual operating cost. Thus even if soda ash make-up
to a system were to increase by 100% over the expected value,
operating cost might only be increased by a factor of 1.02. On
the other hand the environmental consequences of higher sodium
consumption may be significant if the sodium is leachable from
the waste product to the environment.
A logical way to measure sodium consumption is in moles of
Na consumed per mole of sulfur removed by the system. A value of
0.05 moles of Na make-up per mole of sulfur removed (equivalent
to 0.025 moles of Na2C03 make-up per mole of sulfur removed)
appears to be a reasonable design target based on present U.S.
technology. This target is achievable in concentrated alkali
systems burning relatively high sulfur coal (over 3% sulfur) and
having a relatively low oxygen content flue gas. In Japan sodium
make-up is reportedly as low as 0.02 moles Na/mole sulfur removed
for some systems.
Calcium Consumption
A logical way to specify calcium consumption is as calcium
stoichiometry moles of calcium added per mole of sulfur removed
(or collected). A calcium consumption of 0.98 to 1.0 appears
to be a reasonable design target for concentrated double alkali
systems (values under 1.0 are possible due to the additional alkali
added with sodium make-up to the system).
Power Consumption
Design targets for power consumption can be in the range of
1-2% of power plant output without reheat or under 3% even with
50°F of reheat. These figures are based on a system which has a
scrubber pressure drop in the range of 6-8 in. H20 and a scrubber
system L/C ratio of about 10 gal./lOOO ft3. This assumes that the
power plant is equipped with some means of efficient particulate
collection upstream o! the FGD (e.g., an electrostatic precipitator)
411
-------
Low power consumption is a major selling point for double
alkali FGD systems. Some other FGD systems are estimated to run
as high as 10% in power consumption.
Economics
It appears that double alkali systems are now economically
competitive with the "first generation" wet alkali lime/limestone
slurry scrubbing systems. This is especially true in cases where
the lime or limestone system would be required to be equipped with
solids separation equipment (i.e. thickener and filter).
Some of the factors that allow a double alkali system to be less
expensive than a lime/limestone system both in initial cost and
annualized operating cost are:
• Lower scrubber liquid/gas ratio (L/G)
• Lower scrubber nressure drop (AP)
• Simpler scrubber design
- fewer stages
- no slurry in scrubber
• Less exotic materials of construction
• Solid waste with better handling properties
It is estimated that a new (or simple retrofit) double alkali
system could be installed at a 200 Mw or larger plant for a cost of
$50 - 60/kw and operated at 2.5 - 3 mills/kw-hr excluding sludge
disposal (which would vary with the particular application).
These estimates are based on the following assumptions:
200 Mw or larger system
3-4% sulfur coal
80% load factor
Capital charges @15-16% of capital investment, annually
Maintenance @3-4% of capital investment, annually
Operation with two operators/shift
Power @$0.02/kw-hr
Soda Ash @ $65-7O/ton
Lime @ $30-33/ton
It appears that economics of scale for units above 200 Mw in
size are not very significant since larger units would probably
involve the installation of additional modules. There would be
some economy of scale in the savings on design, engineering, site
preparation, etc.
412
-------
THE EPA DOUBLE ALKALI RESEARCH, DEVELOPMENT AND DEMONSTRATION
PROGRAM
EPA has been Involved in the development of double alkali technology
since the Second International Lime/Limestone Wet-Scrubbing Symposium
held in New Orleans in November 1971. Some of. the incentive to develop
this technology stemmed from a paper presented by R.J. Phillips
of General Motors (GM) concerning GM's laboratory and pilot plant
work with a dilute double alkali system which appeared very encouraging
at that time in light of the difficulties which were being experienced
with lime/limestone systems. The development of double alkali technology
by EPA has followed an orderly progression of scale from laboratory
to pilot plant to prototype and finally to a planned full-scale utility
demonstration of the process. Along the way, EPA also embarked on a
program to evaluate a full-scale dilute double alkali system in operation
at a GM industrial boiler system (32 Mw equivalent).
Some initial laboratory work on regeneration chemistry was
done in the EPA laboratories at Research Triangle Park in addition
to an initial feasibility study13 which indicated that double
alkali systems might be somewhat lower in capital and operating
costs than lime or limestone systems under certain circumstances.
After these initial studies, EPA contracted with Arthur D. Little
(ADD to conduct a study of the double alkali process. The scope ot
work included In the initial laboratory and pilot plant Program was
subsequently expanded to include prototype testing at the Scholz Plant
of Gulf Power Company where a 20 Mw FGD prototype system was constructed
by Combustion Equipment Associates (CEA)/ADL for Southern Services. The
construction prototype unit was funded by The Southern Company.
The initial goals of the EPA double alkali program were to:
•Demonstrate reliable system operation
•Demonstrate high S02 removal, 95% desirable
•Demonstrate environmentally acceptable sulfate removal schemes
•Minimize soluble material in disposable waste
•Minimize moisture in disposable waste
•Demonstrate closed-loop operation
•Minimize costs
•Minimize Ca^ concentration in the scrubber
To date a high degree of success with these goals has been
achieved up to the prototype level.
413
-------
Laboratory
Some of the areas of investigation in the laboratory program
are listed below:
•Regeneration of simulated scrubber effluents with lime
and limestone .
•Sulfate removal by precipitation of a mixed crystal containing
CaSOA and CaS03 with water of hydration
•Sulfate removal by reaction of Na2S04, CaS03 and sulfuric acid
•Feasible ranges of sulfate, chloride, magnesium and iron in
solution .
•Settling characteristics of the product solids
•Fixation of double alkali product solids
• Density, comparability, leachability and permeability of
fixed and unfixed solids
Pilot Plant
In the pilot plant both short term and long term runs (v, 5 weeks/
run) have been conducted to examine various modes of double alkali
operation. Some of these modes of operation are listed below:
•Concentrated alkali, sulfuric acid treatment, lime
•Concentrated alkali, two-stage reactor system, lime
•Concentrated alkali, single CSTR reactor, lime
•Dilute alkali w/sulfite oxidation, lime
•Concentrated alkali, multi stage reactor, limestone regeneration
•Extra high concentrated alkali, two stage reactor system, lime
Much of the laboratory and pilot plant work has been reported
by LaMantia, et al.3 in a paper presented at the last EPA Symposium
on FGD in Atlanta.
Prototype
The results of testing to date at the 20 Mw prototype system
at Gulf Power will be reported in a paper entitled, "Operating
Experience — CEA/ADL Dual Alkali Prototype System at Gulf Power/
Southern Services, Inc." by LaMantia, et al.17
GM Industrial Boiler
The GM test program at the 32 Mw equivalent system at the Chevrolet
Transmission plant in Parma, Ohio is being conducted under an agreement
between EPA and GM. The -test program design and some chemical analyses
are being done for GM and EPA under contract by ADL. The FGD system consists
of four Koch tray stainless steel scrubbers, 32 Mw equivalent, and a 40 Mw
equivalent regeneration system consisting basically of two mix tanks and
414
-------
two reactor clarifiers. The boiler system consists of two 60,000
Ib/hr (of steam) and two 100,000 Ib/hr boilers. The GM system
is a dilute alkali system.
The system has been operated intermittently since start-up in
February 1974. Availability data for the system is not easily
defined since there are four boilers and four scrubbers while
the steam demand is frequently less than the capacity of one or
two boilers. Thus, when there is a problem with one of the
systems it can be taken out of service and replaced by one of
the stand-by units.
The GM system cost approximately $3.5 million.
A good account of this system was presented in the last EPA
Symposium of FGD in Atlanta by Dingo and Piasecki. 6
Full Scale Utility Demonstration
The RFP by EPA for a full-scale utility demonstration was
issued in May 1975. Proposals were received in response in
August 1975. At the time of this writing (Feb. 1976), the proposals
are in the final stages of evaluation.
The RFP called for a 100 Mw or larger demonstration of the
technology. The three proposals receiving final consideration range
in size from 192 to 575 Mw. The contract will be let for a four-
phase program consisting of:
(I) Process design and cost estimate
(II) Engineering design, construction and mechanical testing
(III) Start-up and acceptance testing
(IV) One year operation and long term testing
A summary of the design criteria requested for the demonstration
unit is given below:
•Unit will meet all applicable pollution control regulations
-Pulverized coal fired boiler burning 2.5 - 4.5% sulfur coal
•Instrumentation must be provided to allow accurate material
and energy balances at the demonstration plant
•S02 control to below 200 ppm emissions by the scrubber
•Stack gas reheat required
•Chemical make-up requirements requested
-Na 0.05 moles/mole S
-Ca 1.2 moles/mole S
.Acceptable long-range waste disposal plan
We anticipate conducting an extensive test program at the full-scale
facility which would characterize effluents, evaluate chemical consumption,
evaluate sludge disposal, optimize operating conditions and evaluate system
reliability and process economics.
415
-------
SUMMARY OF STATUS OF TECHNOLOGY
Based on the initial performance and reliability demonstrated in
various double alkali pilot and prototype plants, and in an increasing
number of full scale applications in the U.S. and Japan, it appears
that this technology has become a viable technological alternative
throwaway process to the lime/limestone processes.
Performance of double alkali systems with respect to S02 removal
is well established. Over 99% S0£ removal has been achieved with lower
than 10 ppm S02 concentration in the treated flue gas.
Potential environmental problems associated with waste disposal
from these systems may occur due to the preserce of soluble sodium salts
which could cause water pollution problems in the surface and ground
water in the vicinity of disposal sites. A number of techniques to
reduce the soluble sodium salts in the waste product have been tested;
however, it appears that there will inevitably be a higher level of
soluble salts in the waste product from double alkali systems than from
lime/limestone systems. With present U.S. technology it appears that
the incremental amount of solubles in a double alkali waste product
(over that typically present in lime/limestone waste product) is
typically about 1-2% on a dry solid? basis. Development of sludge fixation
technology to change the wet solid waste product to a hard unleachable
solid could conceivably reduce this potential problem.
Actual costs for full scale utility boiler applications are not
available. Estimates for these applications based on the best avail-
able information put double alkali system capital costs in the range
of lime/limestone system costs, at about $50-60/kw capital and
2.5-3 mills per kw-hr operating cost.
In the U.S., development has stressed the use of lime rather than
limestone as the calcium source in all of the full scale industrial
applications and in most of the pilot plant testing in both dilute and
concentrated systems. In Japan, limestone rather than lime is used for
concentrated systems in full scale utility boiler applications apparently
for the operating cost benefit which is derived through the use of the
less expensive regenerant. Also, due to differences in the economies
of the two countries and state-of-the-art of technology, production of
by-product gypsum from the Japanese double alkali systems is prevalent;
416
-------
whereas, production of a throwaway solid waste product is the
general rule with development of the technology in the U.S.
The table presented below shows a total of approximately
3900 Mw equivalent operating and planned double alkali systems
in industrial and utility applications in the United States and
Japan representing 23 applications of this technology. Of these,
approximately 1750 Mw equivalent representing 16 applications
are operational. In the U.S. six operational units approximately
equivalent to 140 Mw are listed. As a standard for comparison
the latest PEDCo Survey of FGD systems in the U.S.1^ indicates a
total of 42,160 Mw representing 109 units of operating and planned
FGD systems. Of these 3828 Mw representing 22 units are operational.
The development of double alkali technology has obviously
stemmed (and benefitted) from both lime/limestone and other soluble
sodium scrubbing technology development. Possibly as a result of
this and certain inherent advantages of soluble alkali scrubbing,
it appears that the reliability established at this point in the
development of double alkali technology is greater than that which
had been established for the lime/limestone systems at a corresponding
stage of development.
417
-------
SUMMARY OF SIGNIFICANT OPERATING AND PLANNED FULL SCALE DOUBLE ALKALI SYSTEMS
System Operator
FMC
Modesto Calif. Plant
Showa Denko KK
Chiba, Japan
Tohoku Electric
Shinsendai Sta. ,
Japan
General Motors
Parma, Ohio Plant
Caterpillar Tractor Co.
Joliet, Illinois
Showa Pet. Chetn.
Kawasaki, Japan
Kanegafuchi Chera.
Takasago, Japan
Poly Plastic
Fuji, Japan
Kyowa Pet. Chem.
Yokkaichi, Japan
Kinuura Utility
Nagoya, Japan
Daishuwa Paper
Fuj 1 , Japan
System ADp_lica t ion
Reduction kilns
Oil-fired elee.
power boiler
Oil-fired
utility boiler
4 coal-fired
InJustrial boilers '
2 coal-fired
industrial boilers
Oil-fired
industrial boiler
Oil-fired
industrial boiler
Oil-fired
Industrial boiler
Oil-fired
industrial boilar
Oil-fired
industrial boiler
Oil-fired
boiler
Vendor or
Developer
FMC
Showa
Denko
Kawasaki/
Kureha
General
Motors
Zurn
Industries
Showa
Denko
Showa
Denko
Showa
Denko/Ebara
Showa
Denko/Ebara
Tsukishii-ia
Tsukishima
Size
(Mw. Equivalent)
10 (Gas Rate-)
30 (Regen.)
150
150
40 (Regen)
32 (Gas Rate)
v 20-30
62
93
65
46
63
85
Active
Alkali
Cone .
Cone .
Cone.
Dilute
Dilute
Cone.
Cone.
Cone.
Cone.
Cone.
Cone .
Calciur. Start-Up
Sources Dite'1
Lin't Dec. 1971
Limestone June 1973
Limestone Jan. 1974
lime Mar. 1974
Lime Oct. 1974
Limestone 1974
Limestone 1974
Limestone 1974
Limestone 1974
Lime 1974
Lime 1974
-------
SUMMARY OF SIGNIFICANT OPERATING AND PLANNED FULL SCALE DOUBLE ALKALI SYSTEMS (Continued)
System Opcra_tcrt^
Firestone
Pottstown, Pa.
Gulf Power Company
Scholz Plant
(Southern Services)
Caterpillar Tractor Co.
Mossville, Illinois
Sikoku Electric Power
Anan, Japan
Sikoku Electric Power
Sakaide, Japan
Kyushu Electric Power
Buzen, Japan
Caterpillar Tractor
East Peoria, Illinois
Central Illinois Publicb
Service - Newton til
Louisville Gas & Electric1'
Cane Run #6
Louisville, Kentucky
Springfield Water Light*5
and Power - Dallman #3
Springfield, Illinois
Caterpillar Tractor
Mapleton, Illinois
System Application
Coal 6. oil-fired
industrial boiler-
demonstration
Coal-fired
utility boiler-
prototype
Coal-fired
industrial boiler
system
Oil-fired
utility boiler
Oil-fired
utility boiler
2 oil-fired
utility boilers
Coal-fired
industrial boiler
Coal-fired
utility
Coal-fired
utility
Coal-fired
utility
Vendor or
Developer
FMC
A.D. Little''
Combust . Equip .
Assoc iates
FMC
Kawasaki/
Kureha
Kawasaki/
Kureha
Kawasaki/
Kureha
FMC
Envirotech
CEA/ADL
FMC
Size
(Mw, Equivalent)
3
20
45-35
450
450
(2-450)
900
100
575
280
192
Active Calciur.i
Alkali Sources
Cone. Lime
Cone. Lime
Cor,.:. Li '.Tie
Cone . Linest oni-
Cone. Li-estoni;
Ccnc. Li.-estone
Cone. Lime
Cone. Line
Cone. Lime
CMC. Line
Start-up
Da t ea
Jar.. 19T5
Feb. 1975
•Jc t . 1 ? 7 3
1975
1975
(May 1977)
(1977)
(1977-78)
(1977-73,)
(1977-78)
Coal-fired
industrial boiler
FMC
100
Cone .
Line
(1978)
Dates in parentheses are projected start-up dates
These units were proposed in response to the RFP for a full scale utility demonstration
-------
BIBLIOGRAPHY
1. Ponder, W. H., "Status of Flue Gas Desulfurization Technology For
Power Plant Pollution Control." Presented at Thermal Power
Conference, Washington State University, Pullman, Washington,
October 4, 1974.
2. Epstein, M., R. Borgwardt, et al., "Preliminary Report of Test
Results from the EPA Alkali Scrubbing Test Facilities at the
TVA Shawnee Power Plant and at Research Triangle Park."
Presented at Public Briefing, Research Triangle Park, North
Carolina, December 19, 1973.
3. LaMantia, C., et al., "EPA-ADL Dual Alkali Program Interim
Results." Presented at EPA Symposium on Flue (las Desulfurlzation,
Atlanta, Georgia, November 4-7, 1974.
4. Johns tone, II. F., et al., "Recovery of Sulfur Dioxide from Waste
Gases." Ind. & Eng. Chem., Vol. 30, No. 1, January 1938,
pp 101-109.
5. Phillips, R. J., "Operating Experiences with a Commercial Dual-
Alkali S02 Removal System." Presented at the 67th Annual Meeting
of the Air Pollution Control Association, Denver, Colorado,
June 9-13, 1974.
6. Cornell, C. F. and D. A. Dahlstrom, "Performance Results on a
2500 ACFM Double Alkali Plant for S()2 Removal." Presented at
the 66th Annual Meeting of A.I.Ch.E., Philadelphia, Pennsylvania,
November 11-15, 1973. Condensed version of the paper appeared
in December 1973 CEP.
7. Kaplan, N., "An EPA Overview of Sodium-Based Double Alkali
Processes - Part II Status of Technology and Description of
Attractive Schemes." Presented at the EPA Flue Gas Desulfurization
Symposium, New Orleans, Louisiana, May 14-17, 1973.
8. Phillips, R. J., "Sulfur Dioxide Emission Control for Industrial
Power Plants." Presented at the Second Internation.il Lime/Limestone
Wet-Scrubbing Symposium, New Orleans, Louisiana, November 8-12, 1971.
9. Brady, J. D., "Sulfur Dioxide Removal Using Soluble Sulfites."
Presented at Rocky Mountain States Section Air Pollution Control
Association, Colorado Springs, Colorado, April 30, 1974.
420
-------
10. Cornell, C. F. , "Liquid-Solids Separation in Air Pollutant Removal
Systems." Presented at the ASCE Annual and National Environmental
Engineering Convention, Kansas City, Missouri, October 21-25, 1974.
11. Selmeczi, J. G. and R. G. Knight, "Properties of Power Plant Waste
Sludges." Presented at the Third International Ash Utilization
Symposium, Pittsburgh, Pennsylvania, March 13-14, 1973.
12. liJ.li.fon, W. , et al., "System Reliability and Environmental Impact
of S02 Scrubbing Processes." Presented at Coal and The Environ-
ineiit, Technical Conference, Louisville, Kentucky, October 22-24,
1974.
1'3. Rochelle, G. T., "Economics of Flue Gas Desu.l fur Iza Lion."
Presented at EPA Flue Gas Desulfurization Symposium, New Orleans,
Louisiana, May 14-17, 1973.
14. "Sulfur Dioxide and Flyash Control", FMC Corporation Technical
Bulletin. FMC Corporation,. Air Pollution Control Operation,
751 Roosevelt Road, Suite 305, Glen F.llyn, Illinois 60137.
15. McClamery, G. G. and R. L. Torstrick, "Cost Comparisons of Flue
Gas Desulfurization Systems." Presented at the EPA Symposium on
Flue Gas Desulfurination, Atlanta, Georgia, November 4-7, 1974.
16. Dingo, T. and E. Piasecki, "General Motor's Operating Experience
witii a Full-Scale Double Alkali Process." Presented at the EPA
Symposium on Flue Gas Desulfurizntion, Atlanta, Georgia,
November 4-7, 1974.
17. LaMantia, C.R., R.R. Lunt, R.E. Rush, T. Frank, N. Kaplan,
"Operating Experience — CEA/ADL Dual Alkali Prototype
System at Gulf Power/Southern Services, Inc." Presented at
the EPA Symposium on Flue Gas Desulf urization, New Orleans,
Louisiana, March 8-11, 1976.
18. Ando, J., "Status of Flue Gas Desulfurization and Simultaneous
Removal of S02 and NOX in Japan." Presented at the EPA
Symposium on Flue Gas Desulfurization, New Orleans, Louisiana,
March 8-11, 1976.
19. Summary Report. Flue Gas Desulfurization Systems. Nov. - Dec. 1975
Prepared by PEDCo Environmental Specialists for U.S. EPA under
Contract No. 68-02-1321.
421
-------
OPERATING EXPERIENCE--CEA/ADL DUAL ALKALI PROTOTYPE SYSTEM
AT GULF POWER/SOUTHERN SERVICES, INC.
Charles R. LaMantia and Richard R. Lunt
Arthur D. Little, Inc.
Cambridge, Massachusetts
Randall E. Rush
Southern Services, Inc.
Birmingham, Alabama
Thomas M. Frank
Combustion Equipment Associates, Inc.
New York, New York
Norman Kaplan
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ABSTRACT
As part of the Dual Alkali Program being conducted by Arthur D.
Little, Inc. for EPA's Industrial Environmental Research Laboratory-RTP,
a one-year test is being conducted on the 20-megawatt dual alkali SO
.control process at Gulf Power Company's Scholz Steam Plant in Sneads,
Florida. The system was developed, designed and installed by Combustion
Equipment Associates, Inc./Arthur D. Little, Inc. for Southern Services,
Inc. Initial startup of the system occurred in February 1975; after
the shakedown period, the EPA test program commenced in mid-May, 1975.
This paper presents a description of the system and its performance
during the first year of operation from the initial startup through
early January 1976, when the boiler was shut down for a scheduled overhaul,
The boiler and the control system are due to be put back in operation
by early March, 1976.
423
-------
ACKNOWLEDGEMENTS
The authors would like to acknowledge several individuals for their
invaluable contributions to the test program for the prototype dual
alkali system.
We greatly appreciate the efforts of Indrakumar Jashnani, Bernard
Jackson, and James Valentine of ADL who helped coordinate and supervise
the test program at Scholz. Steve Spellenberg and Bruce Goodwin of
ADL provided the bulk of the on-site chemical analytical support.
The cooperation of Combustion Equipment Associates, Gulf Power Company,
and Southern Services, Inc. in this prototype test program is also
gratefully acknowledged. In particular, we would like to thank
Richard White of Combustion Equipment Associates, Reed Edwards of
Southern Services, and James Kelly of Gulf Power for their on-site
assistance in the operation of the system. Their continuing concern
and attention to the various aspects of the test program and the
system operation have been valuable contributions to the success of
our overall efforts.
424
-------
OPERATING EXPERIENCE--CEA/ADL DUAL ALKALI PROTOTYPE SYSTEM
AT GULF POWER/SOUTHERN SERVICES, INC.
I. SUMMARY
During the first year of operation after startup, the prototype dual
alkali system operated at an S02 concentration below the design range and
at oxygen concentrations which exceeded the design range until the last
three to four months of operation. Over the course of the year, the
process and equipment performance improved significantly with excellent
system availability during the final operating period from mid-September
1975 through early January 1976. The process has demonstrated very good
reliability, stability of operation, and resistance to upset with no
incidents of scrubber or mist eliminator plugging or scaling, even during
the last three months of operation with no mist eliminator wash.
Under these conditions, the system operation relative to important
performance parameters was as follows:
• S02 Removal—The system generally operated at S02 removal
efficiencies of about 95% and demonstrated the capability for
removal of over 99%. The amount of S02 removed is controllable
by adjustment of scrubber pH.
• Oxidation and Sulfate Formation and Control—By co-precipitation
of sulfate with sulfite, this concentrated dual alkali mode can
keep up with oxidation rates of up to 25-30% in closed-loop
operation.
• Lime Utilization—Lime utilization typically ranged from 95% to
100%. The overall lime stoichiometry was 0.95-1.00 moles Ca(OH)2/
mole of S02 removed from the flue gas.
• Waste Solids Properties—The system produced a washed filter cake
generally containing at least 50% insoluble solids in this low to
medium sulfur coal/high oxidation situation. The waste material,
on the average, contained 3-5% soluble solids (dry cake basis,
Period 2) with soluble solids reduced to the 2-3% range when
washing was well controlled. Under process conditions consistent
with higher sulfur coal operation, the insoluble solids content
of the cake improved markedly with improved washability and lower
soluble solids content.
• Sodium Makeup Requirements—About 0.03 moles Na2C03 are required
per mole of S02 removed as makeup for soluble solids losses in
the washed cake under these process conditions. About an equal
amount of sodium makeup was unaccounted for, probably due to
leakage from the system and errors in soda ash makeup estimates.
An attempt will be made to reduce leakage by maintenance performed
during the shutdown period in January 1976.
fmerally, the system performance has exceeded expectations for these flue
as conditions.
425
-------
It is planned to put the system In operation again by early March 1976
and continue operations through about June 1976. During this next
operating period, high sulfur coal (3.5-4.5%) will be in use at Scholz,
enabling evaluation of the system performance in a high sulfur coal
application.
426
-------
II. INTRODUCTION
The 20-megawatt, prototype dual alkali process was designed and installed
by Combustion Equipment Associates, Inc. (CEA) and Arthur D. Little, Inc.
(ADL) at the Scholz Steam Plant of Gulf Power Company near Sneads, Florida.
This system is one of three "second generation" 20-megawatt, prototype
flue gas desulfurization systems installed at the plant as part of a
technology evaluation program being conducted by Southern Services, Inc.
for The Southern Company (an electric utility holding company including
Alabama Power Company, Georgia Power Company, Gulf Power Company, Mis-
sissippi Power Company, Southern Electric Generating Company and Southern
Services, Inc.).
rhe process was developed and designed jointly by CEA/ADL. Early labo-
ratory research on the process, performed by ADL and sponsored by the
Illinois Institute for Environmental Quality, dealt exclusively with
characterizing the nature of the regeneration reaction. Based upon the
laboratory results, a 2,000 cfm* dual alkali pilot plant was conducted
at ADL's facilities by CEA/ADL and an eight-month test program was con-
ducted to generate the design data for the prototype system. The pilot
system contains the complete dual alkali process loop involving: gas
scrubbing; absorbent regeneration; and solids separation. Results of
the laboratory program and pilot operations for generation of the proj-
totype system design have been reported previously in the literature.
The laboratory and pilot plant investigation of dual alkali technology
has continued at ADL in a program for the U.S. Environmental Protection
Agency's (EPA) Industrial Environmental Research Laboratory at Research
Triangle Park, N. C. The program involves characterization of the basic
process chemistry and the various modes of operation of sodium-based
dual alkali processes. The work covers a wide range of flue gas condi-
tions, liquid reactant concentrations, and process configurations
including the use of both lime and limestone for regeneration of the
sodium scrubbing liquor. The three tasks of this program are described
briefly below. Interim results were presented at the 1974 EPA Flue
Gas Desulfurization Symposium.
In Task I, the ADL Laboratory Program, experiments were performed on the
regeneration of concentrated sodium scrubbing solutions using lime re-
generation, limestone regeneration and sulfuric acid treatment for
sulfate removal. Work was performed at EPA's laboratories at Research
Triangle Park on regeneration using limestone in dilute mode operation.
* Despite EPA policy, certain non-metric units are used in the paper;
however a conversion table is provided at the end of the paper.
427
-------
Results of this prior work have been reported.2 More recent work still
in progress includes a characterization of the chemical, physical, and
crystallographic properties of dual alkali solids, including work on
the nature of calcium sulfate species co-precipitated with calcium
sulfite in dual alkali processing. Work on physical properties includes
evaluation of the mechanical and engineering properties of untreated
dual alkali solids and solids produced from a limited number of mixtures
of dual alkali solids with fly ash, lime and other agents used in the
treatment of solid wastes from conventional lime/limestone flue gas
desulfurization processes. This work is scheduled to be completed and
reported in mid-1976.
Task II, the Pilot Plant Program, was conducted at the CEA/ADL pilot
facility, Cambridge, Massachusetts. Results of short-term pilot opera-
tions using concentrated sodium scrubbing solutions with lime regenera-
tion, limestone regeneration and sulfuric acid treatment for sulfate
precipitation were reported previously.2 Since then, short-term opera-
tions have also been conducted using dilute sodium (Na+ active) scrubbing
solutions with lime regeneration; and additional pilot operations have
been conducted in an attempt to develop a successful dual alkali system
using limestone for regeneration.
From the short-term pilot operations, three of the most promising modes
of dual alkali processes were selected for long-term (three to five
weeks around the clock), closed-loop pilot yuns. These long-term tests
were conducted in the following dual alkali modes:
• Dilute mode using lime for regeneration
• Concentrated mode using lime for regeneration
• Concentrated mode using limestone for regeneration
Work on the dilute and concentrated modes using lime has been completed.
The results indicate that these modes can be operated closed loop with
the following general performance:
• SQ2 removal - 90% or greater
• Lime utilization - 95% or greater
• Waste cake solids content - 45% or greater
• Sodium makeup requirements - less than 0.05 moles Na2C03
per mole of SQz removed
The actual performance of the dual alkali process will vary depending
upon the SQz and oxygen concentrations in the flue gas, the design of
the system and the concentration of sodium solutions used in the process.
Some version of the dual alkali process can generally be designed to
exceed many or all of the above performance characteristics in most
utility applications.
428
-------
Aside from concentration differences, the principal difference in
operating characteristics between dilute and concentrated lime dual
alkali systems is that the dilute systems operate at or near saturation
in calcium sulfate, which requires the use of carbonate makeup to pro-
vide some softening of the regenerated solution prior to recycle back
to the scrubber.
Operations using limestone for regeneration have not been completed.
Recent results have been promising, with many of the performance charac-
teristics approaching those of lime dual alkali systems. The problem
which remains to be solved is the production of solids with good settling
characteristics over a wide range of sulfate, magnesium and iron concen-
trations in the scrubbing liquor.
The results of these long-term runs and recent work on limestone regen-
eration will be presented in' a report scheduled for mid-1976.
Task III, the Prototype Test Program, consists of a one-year test of
the 20-megawatt CEA/ADL prototype dual alkali system at Gulf Power.
The objective of the program is to characterize the important aspects
of the prototype process performance:
S02 removal efficiency
Oxidation and sulfate formation and control
Lime utilization
Waste solids properties
Sodium makeup requirements and degree of closed-loop operation
Process reliability
The prototype system was completed and put in operation by CEA/ADL in
early February 1975. In mid-May 1975, the test program was initiated
by ADL, Southern Services and CEA as part of the EPA/ADL Dual Alkali
Program. The system was operated until early January 1976 when the
boiler was shut down for a scheduled overhaul. The boiler and the
prototype system are due to be put back in operation by early March 1976.
This paper describes the performance of the system over this first year
of operation from startup in February 1975 through early January 1976.
The startup period, including the first three months of operation prior
to the EPA program, has already been reported in detail in the litera-
ture3; for completeness in describing the first year of operation,
information'on the first three months of operation is also included.
429
-------
III. CEA/ADL 20-MEGAWATT DUAL ALKALI SYSTEM
A. PROCESS CHEMISTRY
The CEA/ADL dual alkali S02 control process at Scholz Station is a
sodium solution scrubbing system in which the absorbent solution is
regenerated using lime with the provision for use of limestone. The
use of limestone is contingent upon successful pilot plant tests of a
limestone mode.
The absorption of S02 is accomplished using a solution of sodium sulfite,
sodium hydroxide and possibly some sodium carbonate (makeup) producing
a spent sodium sulfite/bisulfite liquor:
2NaOH + S02 -»• Na2S03 + H20
Na2C03 + S02 •*• Na2S03 + C02t
Na2S03 + S02 + H20
2NaHS0
(1)
(2)
(3)
During absorption, and to a lesser extent through the remainder of the
system, some sulfite is oxidized to sulfate:
Na2S03
1/2 0
(4)
converting an "active" form of sodium to an "inactive" form. Oxidation
in the scrubber is generally a function of the scrubber design, oxygen
content of the flue gas and the scrubber operating temperature. At
excess air levels normally encountered in utility power plant opera-
tions burning medium or high sulfur coal, the level of oxidation is
expected to be on the order of 5-10% of the sulfur dioxide removed.
The scrubber solution is regenerated by reaction with lime (and/or
limestone) which precipitates a mixture of calcium sulfite and calcium
sulfate solids for disposal, as shown by the following overall reactions
(shown for lime):
2NaHS03 + Ca(OH)2 -»• Na2S03 + CaS03 -1/2 H204- + 3/2 H20 (5)
Na2S03 + Ca(OH)2 + 1/2 H20 + 2NaOH -I- CaS03-l/2 H2G+ (6)
Na2SOlt + Ca(OH)2 •* 2NaOH + CaSO^I (7)
430
-------
After regeneration, the solids are separated from the regenerated liquor
and washed. The clear liquor, containing very low amounts of suspended
and dissolved calcium, is returned to the scrubbers. Soluble calcium
levels are generally less than 100 ppm in concentrated dual alkali
processes. Any sodium value lost with the washed waste solids is
replaced by the addition of sodium carbonate to the regenerated liquor;
sodium hydroxide can be used since carbonate softening is not required.
Since sodium sulfate is also reacted with lime in this system to regen-
erate sodium hydroxide, it should be possible to use sodium sulfate as
the sodium make-up source when low to moderate oxidation levels are
encountered.
The system is designed to operate in the concentrated active sodium
mode (active Na+ concentration greater than 0.15M). In this mode,
sulfate removal cannot be accomplished by the precipitation of gypsum
'(CaSOit -2H20), since the high sulfite levels prevent the soluble
calcium concentration from reaching that required to exceed the gypsum
solubility product. However, calcium sulfate (CaSO^) is precipitated
along with calcium sulfite (CaS03-l/2 ^0) in the regeneration reactor,
resulting in a solid solution of the two salts. The amount of sulfate
precipitated in this form is a function of the concentrations of species
in solution and the reactor pH. Under normal operation, with sulfate
levels up to 1.5M SOi^, the system can keep up with sulfite oxidation
rates equivalent to 25-30% of the S02 absorbed without becoming satu-
rated in calcium sulfate.
Additional details of dual alkali chemistry and terminology are given
in recent publications.2 >*+
B. SYSTEM DESCRIPTION
The prototype system at the Scholz Steam Plant is installed on Boiler
No. 1, a 40-megawatt (nominal) Babcock and Wilcox pulverized coal-
fired boiler. The prototype is sized to handle 50% of the flue gas
from the boiler. The boiler has been retrofitted with a sectionalized,
high efficiency electrostatic precipitator capable of 99.5% particulate
removal. Sections of the electrostatic precipitator can be de-energized
to study the impact of particulate input on process operation.
The design basis for the prototype system is given in Table 1; a
schematic flow diagram is given in Figure 1. The system, designed as
a prototype, incorporates a high degree of flexibility aimed at
generating design and operating information for a wide variety of
applications throughout The Southern Company system. Although the
basic mode of operation of the system is a dual alkali process with
lime regeneration, the system was designed to accommodate limestone
regeneration and a combination of regeneration with limestone and lime.
The system was also designed to enable operation as a direct limestone
or lime scrubbing system. As a consequence, the system contains equipment
and piping in excess of that required to operate a dual alkali system on
a boiler already equipped with adequate particulate control.
431
-------
TABLE 1
DESIGN BASIS
Flue Gas Inlet
Flow Rate (acfm)
Temperature (°F)
02 Concentration (% volume, dry)
Particulate Loading (gr/scf, dry)
S(>2 Concentration (ppm, dry)
75,000
275
6.5 (max.)
0.02 (precipitator
energized)
1800-3800
Design Performance
S02 Removal (% reduction)
Maximum S02 Removal Rate (Ib/hr)
Particulate (gr/scf, dry)
Power Consumption (% power output)
with venturi, full spray absorber pump
requirements
without venturi, with tray absorber pump
requirements
90 (min.)
1530
0.02 (no increase
in loading)
2.5-3.0
1.0-1.5
432
-------
Scrubbed
Gas
Vacuum Water Recovery. (Filter
Na2CO3
Silo
Solids
Figure 1, CEA/ADL dual alkali process flow diagram - Scholz Station.
-------
The venturi scrubber is included in the system to investigate the opera-
tion of that type of scrubber in a dual alkali mode (or for direct lime
or limestone scrubbing) when simultaneous particulate and S02 removal
is required. The second scrubber, an absorption tower, can be operated
as a tray scrubber, as in a dual alkali system where simultaneous
particulate control is not required; or as a spray tower, for direct
lime or limestone scrubbing, again without simultaneous particulate
control. In addition to this redundancy in scrubbers, sufficient pump
capacity is provided to operate the venturi at an L/G of 25 gallons/
Macf of saturated gas and for an absorber L/G of 60, for operating in
a spray tower configuration.
An additional storage silo (for limestone), a mix tank and other assorted
tanks, pumps, controllers and piping were included in the system to
accommodate the high degree of flexibility desired in the prototype
system. To date, except when limestone has been erroneously delivered
to the system instead of lime, the system has been operated only in
the lime dual alkali mode as given in Figure 1.
In the lime dual alkali mode, the system design is based upon removal
of at least 90% of the S02 in the flue gas for medium to high sulfur
fuels (up to 5% sulfur) encountered in The Southern Company system.
With the precipitator energized, the system was specified not to
increase particulate loadings in the scrubber outlet above those in
the inlet flue gas.
The power consumed by the system (not including reheat oil) is equiva-
lent to 2.5-3.0% of the power generated by the unit in producing the
flue gas load to the system, with the system operating at full fan
capacity (full gas flow at a system pressure drop of 20 inches water)
and at full venturi and absorber liquid recirculation capacity. Cor-
recting for the excess fan and pump capacity, the power consumed by
the equipment actually required in this application (tray tower at an
L/G of 5-10) is roughly 1.0-1.5% of the power generated at the design
conditions. In a full scale system designed for S02 removal only,
the power consumption should be in this range.
The dual alkali system can be conveniently broken down into three
process subsystems: gas scrubbing; absorbent regeneration; and solids
dewatering. The design and operation of each of these subsystems is
discussed in the following sections.
1. Flue Gas Scrubbing
The gas scrubbing system consists of a variable throat, plumb-bob type
venturi scrubber followed in series by an absorption tower. Treated
flue gas flows through both scrubbers. Each of these scrubbers is
equipped with a removable liquid entrainment separator, an enclosed
recycle tank to contain the scrubbing liquor, and two recycle pumps
(one operating and one spare). The venturi can be used for absorption
434
-------
and/or particulate control and can be operated on a separate liquor loop
from the absorption tower or in series with the absorber liquor loop.
The absorption tower can be operated as a tray tower (with up to four
trays), as a. spray tower, or as a de-entrainment separator.
Gas is pulled from the exit of the electrostatic precipitator and
forced through the scrubbing circuit using the booster fan provided
with the dual alkali system. The fan and motor have been designed
fop a total system pressure drop of 20 inches H20 at maximum flow.
Under normal conditions with the precipitator in service, the venturi
was used only for gas saturation with some S02 removal; a pressure drop
of roughly 5 to 10 inches of water was maintained across the venturi.
Gas from the fan enters the venturi scrubber, flows downward over the
wetted approach section and enters the high velocity venturi throat.
Recycled scrubbing liquor also enters the top of the scrubber through
tangential pipe inlets and through a number of vertical bull nozzles
equally spaced around the center of the venturi.
After passing through the throat, the flue gas and scrubbing liquor
continue downward through the internal downcomer; the liquor is
collected at the bottom of the scrubber in t ie ->.nternal recycle tank.
The flue gas makes a 180 degree upward turn and contacts the chevron-
type entrainment separator (when used). The entrainment separator is
equipped with wash sprays above and below which can be operated con-
tinually or on a sequential timing cycle. Either scrubbing liquor or
fresh water can be used for washing. During dual alkali operations,
this entrainment separator system was completely removed.
Following the venturi scrubber, the saturated flue gas enters the
bottom of the absorption tower. This tower has been designed to
operate either as a spray tower, with one or two sets of sprays, or
as a tray tower, with from one to four trays. The bottom tray is
equipped with a spray underneath to wet the bottom side of the tray.
Gas passes upward through the trays and then through a final de-
entrainment separator, with a wash system similar to the venturi
demister wash. The wash system was used initially, but found to be
unnecessary and its use discontinued.
The clean flue gas leaving the tower is finally reheated by the
injection of hot gas from a burner fired with No. 2 fuel oil before
being discharged through the stack on top of the absorber. Oil-fired
reheat was specified by Southern Services for the prototype systems
in order to reserve steam for power generation.
Regenerated absorbent solution, containing sodium hydroxide, sodium
sulfite, sodium sulfate and some sodium carbonate, is fed to the top
tray of the absorber. The solution flows countercurrent to the gas
through the tray system (two trays) and is collected at the bottom
of the absorber in the internal recycle tank. This collected liquor
supplies solution both for spraying the bottom tray and for recirculation
435
-------
to the top tray, as needed for pH control in the scrubber or to maintain
liquor flow across the trays. The pH in the combined stream of recircu-
lated liquor and freshly regenerated feed ranged from about 6 to 11.
A bleed from the collected tray tower bottoms (pH roughly 5-7) is sent
forward to the venturi recirculation loop for additional S02 removal.
In the venturi, the gases with the highest S02 concentration are con-
tacted with the most acidic liquor. A continuous bleed stream is
drawn from the venturi recirculation loop and is sent to the absorbent
regeneration system. The pH of the bleed stream normally ranged from
4.8 to under 6.0.
2. Absorbent Regeneration
Spent scrubber solution from the venturi recirculation line is bled
to the regeneration reactor where it is reacted with hydrated lime.
A full-scale system would normally incorporate a lime slaker and a
slaked lime slurry tank; hydrated lime was used in this smaller
installation for simplicity. The normal mode of operation was to feed
dry lime directly to the reactor; however, provisions have been made
to feed slurried hydrated lime as well as either slurried or dry
limestone. The rate of the lime feed will in most cases be set accord-
ing to the boiler load and coal sulfur content and adjusted on the pH
of the reactor effluent liquor.
The lime neutralizes the bisulfite acidity in the scrubber bleed and
further reacts with sodium sulfite and sodium sulfate to produce
sodium hydroxide. These reactions precipitate mixed calcium sulfite
and sulfate solids resulting in a slurry containing up to 5 wt %
insoluble solids. The regeneration was normally carried out to a pH
of between 11.0 and 12.5.
The regeneration reactor system consists of a multistage design,
patented by CEA/ADL. The system at Scholz consists of two reactor
stages in series: a short residence time reactor followed in series
by a longer residence time reactor, designed for reactant residence
times of 5 minutes and 35 minutes respectively at the design S02
removal rate (185 gpm flow rate through the reactor system). This
reactor system design has been shown to generate clusters of sulfite/
sulfate crystals which are generally spherical rather than the needle-
like or platelet crystals generally associated with calcium sulfite
precipitation. These crystal clusters have good settling, filtration
and washing properties and can be generated in the reactor system
over a wide range of flue gas and process conditions.
436
-------
3. Solid/Liquid Separation and Solids Dewatering
Slurry from the regeneration reactor system is fed to the center well
of the slurry thickener. The 40 foot diameter thickener is sized to
handle the solids produced from the treatment of flue gases from the
full 40 megawatts of Boiler No. 1.
The thickened slurry from the bottom of the settler is sent to a rotary
drum vacuum filter. The solids content of the underflow was maintained
below 30% to ease underflow pumping. The slurry is recirculated past
the filter in a recycle loop that returns the slurry to the solids zone
in the settler. The feed to the filter is drawn as the bleed from this
recirculation loop. The filter surface area is 75 ft2.
The solids cake is washed on the filter using two or three water sprays
arranged in series. This wash removes a large fraction (up to 90%) of
the occluded soluble salts from the cake and returns these salts to
the system, thereby reducing sodium losses and minimizing sodium car-
bonate makeup.
Solids from the filter are retained in a solids storage enclosure under
the filter from which they are loaded into dump trucks for transfer to
the disposal area. The mixed filtrate and wash liquor from the filter
are returned to the thickener.
Makeup sodium carbonate may be fed to the thickener in order to allow
easy removal of any CaC03 precipitated. This carbonate is not intended
for use as a softener, since soluble calcium concentrations in the
regenerated liquor generally will run less that 100 ppm. However,
some CaCOs precipitation will occur due to its very low solubility
limits. The amount of CaCOs formed will be very small, a few hundred
Bounds per day at design conditions, which is less than 0.5% of the
total cake produced.
Clear liquor overflow from the thickener is collected in the thickener
told tank which acts as surge capacity for the absorbent liquor feed
to the scrubber system. Water can also be added to this hold tank to
lake up for the difference between total system water losses (evapo-
ration and cake moisture) and total water inputs from other sources
{sodium makeup solution, pump seals, lime feed, cake wash, and
demister wash).
!he disposal area for the dual alkali waste cake is a one-acre pit
(450 ft x 100 ft) approximately 12 ft deep. The bottom and sides of the
fit are lined with a layer of clay covered by a double thickness of poly-
ethylene liner. The polyethylene is reinforced between the sheets with
tmesh of nylon fibre. The floor of the pit slopes to a single collec-
tion drain constructed of PVC from which leachate is discharged to the
teh pond. On top of the polyethylene liner is a two-inch layer of sand
Aich gradually changes to gravel near the drain.
437
-------
IV. OPERATING HISTORY
A. GENERAL OPERATING CONDITIONS
The prototype system commenced process startup on February 3, 1975 and
was shut down on January 2, 1976 when the boiler was taken off the
line for a scheduled overhaul. During this time the system was opera-
ted approximately 4,700 hours with shutdown periods of varying length
for system maintenance and modifications and during periods when the
boiler was'taken off the line. Throughout the eleven months, the
system treated flue gas with lower sulfur dioxide concentrations and
higher oxygen concentrations than the range for which the system was
designed. This represented a difficult test for the prototype system.
In sodium-based dual alkali systems for a given scrubber design and
soluble solids level, the rate of sulfite oxidation (moles per unit
time) is a strong function of the oxygen concentration in the flue gas
and is relatively independent of the rate of sulfur dioxide removal
(moles per unit time). As oxygen concentrations increase and SC>2 con-
centrations decrease, a higher percentage of the S02 removed from the
flue gas is converted to sodium sulfate rather than sodium sulfite/
bisulfite in the scrubber liquor. This higher percentage oxidation
requires an increase in precipitation of sulfate relative to sulfite
and a higher calcium sulfate content in the precipitated calcium
sulfite/sulfate to enable closed-loop operation with no intentional
purging of sodium sulfate.
Over the eleven-month period, sulfur dioxide concentrations in the flue
gas averaged between 1,100 and 1,200 ppm, compared with a minimum design
concentration of 1,800 ppm. The actual concentrations varied over a
range of from 600 to 1,600 ppm, often fluctuating daily and even hourly
over the entire range of concentrations. Oxygen concentrations in the
flue gas entering the scrubber ranged from 5.0% to 10.5% (by volume
equivalent to 30-100Z excess air) with oxygen levels increasing with
decreasing boiler load (higher excess air). During the first half of
the year, air leaks in the boiler combustion air preheater and coal
feed tubes contributed from 0.5 to 2.0 volume % to the oxygen concentra-
tions. Oxygen levels in the flue gas were reduced from the 7-10% range
down to the 5-7% range in September after air preheater repairs and when
burner box pressure was brought under better control.
This general change in flue gas oxygen concentration and changes in
the system operation made at about the same time logically led to the
division of this first year of operation into two discrete operating
periods described in the next section.
438
-------
B. DESCRIPTION OF OPERATING PERIODS
1. Operating Period 1—Startup and Initial Operations
This first operating period extends from February 3, 1975 through
July 18, 1975, when the system was shut down for a two-month period
for modifications, repairs and to await receipt of replacement parts.
During this operating period, commencing on initial process startup,
the system was operated approximately 2,600 hours, 83% of the time
that the boiler was in operation. Almost the entire system down-time
of 518 hours relative to the boiler is accounted for by a three-week
shutdown (491 hours) in mid-April, after the start-up period, for
necessary maintenance and'equipment modifications and adjustments prior
to the start of the formal EPA test program in mid-May. Equipment
changes are described in Section VB.
Table 2 contains a summary of the fuel and flue gas characteristics
encountered during this operating period. The boiler was fired with
a combination of several medium and low sulfur coals ranging in sulfur
content from 0.9 to 2.2 wt %. The weighted average sulfur content of
the coal as burned was 1.6 wt %, producing an average S02 level in the
flue gas of about 1,050 ppm. Since it was not possible to segregate
and selectively fire the different coals, the S02 levels in the flue
gas fluctuated daily and often hourly from a low of 600 ppm to a high
of 1,550 ppm.
The flue gas load to the prototype system was kept within a range
equivalent to 15-22 Mw. However, the boiler load varied from 15 mega-
watts to 40 megawatts with an attendant variation in excess air rates.
In addition to the unfavorable process conditions resulting from low
sulfur coal and the high oxygen concentrations, there were frequent
process upsets. These upsets included occasional carry-over of fly
ash with the flue gas and the inadvertent contamination of the lime
supply with limestone (caused by a mix-up in the lime and limestone
deliveries to the prototype systems at the plant). While the presence
of fly ash in the system caused little or no effect on operation or
performance, the limestone did produce temporary changes in the process
chemistry—particularly when a significant fraction of the calcium feed
to the reactor was limestone.
The limestone usually appeared mixed with lime at levels of up to 50%
of the total feed; however, on a number of occasions (for periods lasting
up to one day) the feed to the reactor system was pure limestone. During
periods when only limestone was fed to the reactor, the pH in the reactor
system dropped, SO* removal decreased somewhat and waste cake properties
deteriorated slightly. Normal process conditions were re-established
after the limestone passed through the system.
439
-------
TABLE 2
SUMMARY OF SYSTEM INLET CONDITIONS
Operating Period 1 (2/75 - 7/75)
Range Average
Coal Fired:
Sulfur (wt %) 0.9-2.2 1.6
HHV (Btu/lb coal) 11,500-13,000 12,000
Chloride (wt %) 0.05-0.10 0.08
Inlet Gas:
Gas Load Treated (Mw equiv) 15-22 17
S02 Level (pptn-dry basis) 600-1550 1050
02 Level (Z dry vol.) 5.0-10.5 7.5
440
-------
There was a general rise in the flue gas oxygen concentration over the
operating period due to worsening leakage in the boiler air preheater
section. By mid-July oxygen concentrations were in the 8-10% range,
with dilution of the already low S02 levels down to about 850-950 ppm.
At the resultant high oxidation rates (30-50% of S02 removal), active
sodium levels were inadvertently allowed to drop below 0.15M, into the
range of dilute mode dual alkali operation. Under these conditions,
soluble calcium levels in the regeneration system rose and the regen-
erated liquor became saturated in calcium sulfate, resulting in some
gypsum scaling of the piping at the outlet of the regeneration reactor.
At this same time, mechanical problems in the scrubber system (failure
of two control valves and a pin hole leak in the absorber recycle tank)
required that the system be shut down for repairs. After shutdown it
was decided to await the repair of the preheater leak rather than attempt
to continue operating at the low SC>2 concentrations and increasing oxygen
concentrations.
Based upon this experience, it was decided that in future operations
the active sodium concentration would be maintained well above the 0.15M
level by operating with a slightly higher total sodium concentration in
the system. This would allow the sodium sulfate concentration to increase
in the system during periods of increasing rates of oxidation. Sulfate
levels would continue to rise relative to the sulfite until the rate of
sulfate precipitation as a calcium salt equaled the rate of oxidation.
This can be accomplished without reverting to dilute mode operation at
oxidation rates up to about 25-30% of the S02 removal rate.
The system was shut down from mid-July until mid-September. Repairs
and revisions made during this period were of a mechanical nature rather
than involving process changes and are discussed later. About two
thirds of the shutdown period was to await replacement parts for the
valves that had failed.
2. Operating Period 2—Low to Medium Sulfur Coal Operation
The system was put back in operation on September 16, 1975. From mid-
September to mid-October repairs were made to the air preheater during
boiler outages and adjustments were made in the boiler operation, re-
ducing flue gas oxygen levels down to the 5-6% range. For the remainder
of the test period through January 2, 1976, oxygen concentrations were
generally kept in the 5-6% range. As shown in Table 3, S02 levels were
slightly higher than those encountered during the first operating period,
with an average level of about 1,200 ppm. Similarly, the gas load to
the system was in the same range as that for Period 1. Active sodium
concentrations were maintained in a range for concentrated mode opera-
tion and continuing improvements were made in the mechanical performance
of system components, particularly the filter. One unprogrammed period
af regeneration with limestone occurred in this interval.
441
-------
TABLE 3
SUMMARY OF SYSTEM INLET CONDITIONS
Operating Period 2 (9/75-12/75)
Coal Fired:
Sulfur (wt %)
HHV (Btu/lb coal)
Chloride (wt %)
Range
1.5-3.1
11,900-14,100
0.02-0.14
Average
2.1
13,000
0.08
Inlet Gas:
Gas Load Treated (Mw equiv)
SO2 Level (ppm-dry basis)
02 Level (% dry vol.)
16-19
800-1700
4.5-9.5
18
1220
6.0
442
-------
During this period, until shutdown on January 2, 1976, the prototype
system operated approximately 2,100 hours, 97% of the time the boiler
was in operation.
C. AVAILABILITY—FIRST YEAR PROTOTYPE SYSTEM OPERATION
The objective of the installation of the prototype system was to test
the process chemistry and design on this relatively small scale in
order to evaluate the viability of the process technology. While the
reliability of the process was a principal concern, the 20-megawatt
system was not intended to be a demonstration unit nor a test of the
ultimate availability of such systems when applied full scale.
The system design was based upon scale-up by a factor of about 40
from the CEA/ADL dual alkali pilot plant. Although the prototype
system contains spare pumps in the scrubber and regeneration areas,
other key elements of the system such as the reactor system, lime feed
system and filter station were not spared in this prototype application.
Hultiple installations and/or spare capacity for these critical items
would normally be incorporated in full-scale applications. In spite
of these considerations and the difficult flue gas conditions encountered
during the first operating period, the availability of this prototype
unit from initial startup through the first year of operation is impres-
sive.
Boiler and prototype system operating hours are graphically displayed
by month in Figure 2 and summarized by operating period in Table 4.
The availability during the first operating period, 83%, includes the
initial startup of the system and the three-week shutdown period for
system adjustments and modifications prior to the start of the EPA
test program. Such adjustments would normally be expected after process
startup. During the second operating period, system availability exceeded
97%.
During the two months between operating periods, the system was down
for maintenance and awaiting receipt of spare parts and repair of
air leakage at the boiler air preheater. Including this period in the
availability calculation for this first year yields an overall system
availability of 69.3%. Exclusion of this interim period from the cal-
culation yields availability during operating periods of 89%.
443
-------
Operating
Hours
800
600
400
200
| Boiler Operating Hours (Available to Dual Alkali System)
Q Dual Alkali System Operating Hours
88
1^"
S «
a
fv
SS 2?
l_ *-
—
-
•
X
j
/
•
r
J
X
X
X
X
> i
8
p
/
/
/
/
{
g
en
^
^
/
/
/
/
•» 1
;
J
'
;
^
;
/
i
i
p
|
^ CMCM ""
5
g
E is
0
»
T
t
i
t
f
\
/
X
X
1
7
r
/
/
i
\
/
t
?
\
',
\
',
',
\
-
_
_
i i i i
2/75 4/75 6/75 8/75 10/75 12/75 2/76 4/76
Month of Operation
Figure 2, Availability of the CEA/ADL dual alkali system at Scholz.
-------
TABLE 4
Operating Period 1
PROTOTYPE AVAILABILITY SUMMARY
Dates
2/3/75-7/18/75
Operating Hours
Prototype Boiler5
2591 3109
Availability
Prototype Hours
as percent of
Boiler Hours
83.3%
Interim Period
Operating Period 2
TOTAL YEAR
TOTAL OPERATING
PERIODS
7/18/75-9/15/75
9/16/75-1/2/76
0
2153
4744
4744
1529
2213
6851
5322
0
97.2
69.2
89.1
Hours boiler available to dual alkali system.
445
-------
V. SYSTEM PERFORMANCE
A. EQUIPMENT PERFORMANCE
The equipment performance in terms of overall reliability is reflected
in the system availability and the level of maintenance required.
Even though the availability of the system has been very good, there were
numerous equipment and instrumentation problems. To a large degree, these
problems were mechanical in nature and generally reflected equipment
design or fabrication oversights commonly associated with a first-of-a-kind
prototype system. There were only a few problems encountered that reflected
process chemistry, and these required simple operational and/or equipment
adjustments.
The appendix contains a detailed list of mechanical equipment and
instrumentation problems encountered since startup. Included in the list
are a number of problems of the type that would normally be expected to
occur during startup or items normally requiring maintenance. The fact
that some of the problems shown were not immediately corrected is an
indication that they were either of minor significance or did not cause
important operational problems. The more important problems are discussed
below.
»
1. Equipment Problems
The mechanical equipment problems and maintenance items of primary
concern are discussed below.
Filter. The filter was the largest source of problems in the prototype
system. Because of anticipated corrosion problems associated with high
chloride levels, a large part of the filtration equipment was fabricated
out of fiberglass, which is not as sturdy as stainless steel and more
prone to failures at stress points in the construction. Filter problems
during Period 1 included:
• erosion of the fiberglass scraper blade, resulting in jagged
edges which tore the cloth
• erosion of the bridge valve due to solids carried through cloth
holes
• loss of vacuum due to cracks in the internal drum trunnion tubes
• cracking of the plastic caulking strips, allowing retention ropes
to loosen and releasing the cloth panels
• failure of the fiberglass rocker arm used to agitate the slurry
in the filter tub
446
-------
Modifications were made to the filter during the latter part of
Period 1 and the early part of Period 2. These modifications included
replacement of the scraper blade with one fabricated out of stainless
steel, reinforcement of the rocker arm with stainless steel plates, and
the design of a new method for retaining the filter cloth panels. These
changes along with regular inspection and maintenance and increased
familiarity of Gulf Power personnel with the equipment have significantly
improved filter performance.
Reactor System. Two mechanical equipment problems were encountered
in the reactor system operation: plugging of the dry lime feed chute and
solids buildup in the first reactor. The chute plugging problem was
caused by hot vapors drifting up the chute entrance port and wetting the
lime. The problem was resolved by installation of a vibrator on the chute
with provision for injecting air to prevent vapors from rising into the
chute.
The solids buildup in the first reactor was related to the use of dry
lime feed (as opposed to slurry lime), feed chute location, and the poor
agitation in the first reactor. These caused a deposition of solids both
above and below the liquid surface, particularly in the area near the feed
chute. A simulation of the conditions in this reactor at the CEA/ADL pilot
plant in Cambridge has been successful in confirming the source of the
solids buildup. A new first reactor has been designed and fabricated, and
will be installed prior to restarting of the system. The lime slurry feed
system is also being activated as an alternative to the feeding of dry
lime. Feeding a slurry of slaked lime would be the normal practice in
a full-scale application.
Scrubber. The rubber linings in the bleed control valves on both
the absorber and venturi failed during operations in Periods 1 and 2.
These failures may have been due to throttling to control flow. These
valves were originally sized to accommodate direct slurry scrubbing at
much higher flow rates. These rubber lifted valves are scheduled to be
replaced with stainless steel before the system is put back into operation.
Leakage of solution through the venturi and absorber pump seals has been of
concern primarily because it represents a loss of sodium from the system.
Tests were conducted by Southern Services and Gulf Power personnel to deter-
mine the size of the leaks under various pump packing conditions and to
determine how best to eliminate or minimize the leaks. A Teflon impregnated
packing was installed just prior to shutdown and it appears to have all but
eliminated the leaks. However, further testing with this packing is required.
A slight corrosion and ash buildup on the fan during extended shut-
downs was diagnosed to be due to flue gas leaks through the isolation
damper. Since the fan is located upstream of the system it is constructed
out of carbon steel and significant corrosion was not expected. The
corrosion and ash buildup was generally minor and the fan rotor was easily
cleaned and rebalanced prior to startup. Correcting the leak through the
isolation damper by installing a new damper or use of an air seal system
was considered to be unwarranted.
447
-------
Thickener. The major problems with the operation of the thickener
were plugging of the underflow lines (Period 1) and leaking of liquor
through the bottom due to lining failure (Period 2). Plugging of the
thickener underflow lines was partly due to the design of the underflow
piping, and partly due to the frequent downtime on the filter which
allowed thickener underflow slurry concentrations to exceed 30% solids.
Redesign of the underflow lines using flexible piping and adjustments to
the operational procedures to maintain the underflow slurry concentration
in the 15% to 25% range (by back-flushing with clear hold tank liquor to
dilute the underflow when necessary), have effectively eliminated underflow
plugging problems.
The leak that developed in the bottom of the thickener during the
early part of the second period of operation was small but grew worse
throughout Period 2. After shutdown in January all the solids were pulled
from the thickener and the liquor drained. Inspection of the inside of
the thickener showed sections of lining to be failing due to poor curing
or poor application by the lining supplier. The leaks have beeti located
and patched, and the sections of failed lining are being replaced.
Interestingly, when the thickener was drained, pieces of rope and
a flattened paint can were found in the area of the cone. These items
undoubtedly contributed to difficulties with the thickener underflow
system. A piece of rope had previously been extracted from a seized
underflow pump during operation i'l Period 2.
2. Instrumentation
Instrumentation problems primarily involved the pH units, level
transmitters, and soda ash feed solution control system. Other instru-
mentation problems occurred, but for the most part, these were minor.
Flow-through pH units were originally installed throughout the system.
The piping for these units had a tendency to plug and the electrodes coat
with a fine film, causing a drift in pH readings. In the case of the pH
unit in the reactor system, the lines and probe chamber plugged completely
with solids at low slurry flow rates and the probes eroded away at high
flow rates. The flow-through unit on the reactor system was replaced with
an immersion-type unit fitted with a sonic cleaner midway through Period 1.
This has proved to be much more reliable; however, there have been problems
keeping the sonic cleaner mounted on the probe casing. The unit now
requires routine checking and recalibration about every week to two weeks.
The take-off lines for the flow-through units on the scrubber system
have been or are now being relocated. Also, somewhat higher flow rates
will be maintained through the probes to prevent plugging of piping.
Since close pH control is not required, the problems with the pH units
have not been critical. Also, the scrubber system can be operated on
either the venturi bleed pH or outlet S02 (or for that matter, inlet S02
448
-------
and feed forward flow) with only occasional checks of the bleed pH to
verify the SC>2 monitor. In fact, during a few weeks in the early part of
Period 1, when all pH units and S02 monitors were out of service (due to
delays in obtaining replacement parts), the system has been successfully
operated by taking hourly pH readings with a portable pH unit.
The level transmitter/controllers on a few tanks have been unreliable.
The particular type of unit could not be completely serviced in the field.
Thus, there were periods when inaccurate level indication and/or control
have caused operational problems. These units have been replaced.
Similarly, the soda ash solution feed control system has also been
unreliable. This, in combination with occasional plugging of the feed
gate on the dry soda ash feeder, has made it difficult to close the
overall material balance on sodium. During the last half of the second
operating period, the sodium makeup rate was determined using frequent
checks on the specific gravity of the soda ash solution and recalibration
of the flow indicator along with the inventory of soda ash silo and the
quantity of soda ash delivery. Inventory and delivery Information alone
would not have been sufficient on the short-term due to the large storage
capacity in the silo and the low soda ash feed rates. The soda ash feed
control system is now being revised.
B. PROCESS PERFORMANCE
The general process conditions for the two periods of operation are
summarized in Tables 5 and 6. During Period 1 (February 3—January 18, 1975)
the inlet S02 concentration ranged from 600 to 1,550 ppm and averaged
about 1,050 ppm. Flue gas oxygen levels varied from 5% to about 10% and
typically fluctuated between 6.5% and 8.5%.
The system operating philosophy during Period 1 was to maintain the
active sodium concentration in the system at or above 0.2M Na"*" and to
maintain a total sodium concentration of about 2.0M. This resulted in
a fairly stable operation until the end of June, when the inlet S02 level
fell to 850-950 ppm and oxygen levels rose to 8.5-10.0%. Under these
conditions, without an increase in the sodium carbonate feed rate, the
system chemistry drifted into the range of a dilute dual alkali mode with
active sodium concentrations dropping to below 0.15M Na+. Except for the
two-week period of dilute mode operation in July, the thickener liquor
typically contained 0.20-0.25M active sodium, 0.7-0.9M sodium sulfate,
and 0.10-0.15M sodium chloride (4,000-5,500 ppm Cl~).
At the start of Period 2, it was decided to maintain the active sodium
concentration in the system above 0.3M Na+ and to allow the sodium sulfate
concentration to fluctuate to any level necessary to maintain an equili-
brium between sulfite oxidation in the system and sulfate precipitation.
This would prevent any possibility of deterioration of the system chemistry
to that of a dilute mode.
449
-------
TABLE 5
SUMMARY OF SYSTEM OPERATING CONDITIONS
Period 1
(2/3 - 7/18/75)
Range
Inlet Gas:
Gas Load Treated (Mw equiv.)
S02 Level (ppm - dry basis)
02 Level (Z dry vol.)
Regenerated Liquor Composition:
pH
SO,,* (M)
Cl~ (ppm)
1 1
Ca (ppm)
15-22
600-1,550
5.0-10.5
Range3
10-12.6
0.10-0.35
0.60-1.05
3,000-7,000
20-800+b
Average
17
1,050
7.5
Typical3
11-12.5
0.2-0.25
0.7-0.9
4,000-5,500
50-200
^or March - July 1975, not including periods when limestone fed to reactor.
bCa++ levels above 250 ppm occurred during the period just prior to and
during the dilute mode operation.
450
-------
TABLE 6
SUMMARY OF SYSTEM OPERATING CONDITIONS
Period 2
(9/18 - 12/25/75)
Range
Inlet Gas:
Gas Load Treated (Mw equiv.)
S02 Level (ppm - dry basis)
02 Level (% dry vol.)
Regenerated Liquor Composition:
pH
Na+ ^ (M)
act
SO^" (M)
Cl~ (ppm)
I I
Ca (ppm)
16-19
800-1,700
4.5-9.5
Range3
10-12.8
0.25-0.6
0.6-1.05
1,000-2,100
30-160
Average
18
1,200
6oO
Typical3
11-12
0.3-0.35
0.8-1.0
1,300
70
For October - December 1975, not including periods when limestone fed
to reactor.
451
-------
The flue gas conditions during Period 2 improved over those experienced
in Period 1, particularly in regard to the oxygen levels. Inlet S02 concen-
trations ranged from 800 to 1,700 ppm and averaged about 1,200 ppm. Flue
gas oxygen levels generally were controlled to 5-6% with only infrequent
excursions above 6.5%.
At these flue gas conditions, with the active sodium concentration
maintained above 0.3M, the sodium sulfate concentration typically ran
0.8-l.OM, and the chloride concentration fluctuated in the 0.03-0.06M
range (^1,300 ppm Cl~). Routine analyses of coal samples performed for
the plant from both Periods 1 and 2 show about the same chloride content,
approximately 0.08 wt % on the average. However, the accuracy of these
analyses is only about ±0.05 wt %.
The various aspects of the process performance during the two operating
periods are discussed in the following sections. The discussion has been
broken down into six areas:
S(>2 removal
Lime utilization
Oxidation and sulfate control
Waste cake properties
Sodium makeup, and
Process reliability
Complete material balances could be closed only during selected intervals
in each operating period because of the frequent contamination of the lime
supply with limestone, fluctuating flue gas and operating conditions, and
occasional instrumentation problems. However, sufficient analytical and
operational data exist throughout both periods (even when material balances
were not complete) to allow thorough characterization of the system
operation and performance on the low and medium sulfur coal fired during
the first year.
1. S02 Removal
The scrubber system was operated using two different configurations
for S02 removal: the venturi and absorber together in series (with two
trays) and the venturi alone. In the latter configuration the trays were
not removed from the absorber; rather, the regenerated liquor feed to the
top tray was diverted either to the absorber recycle tank (from which it
was transferred to the venturi) or bypassed directly to the venturi
through a line Installed in May 1975. (The bypass line was not in the
original design, since the operation on low sulfur coal was not anticipated.)
The recycle flow to the top tray and the flow to the spray on the underside
of the bottom tray were both discontinued; however, the absorber pumps
were maintained in operation to transfer liquor collected in the absorber
tank to the venturi. In order not to overly back-pressure the absorber
pumps, a recycle was maintained through an open spray header during
intervals when there was no feed to the trays.
452
-------
During Period 1, both operational configurations were used at different
times. However, during Period 2, only the combined venturi and absorber
configuration was used.
The SC>2 removal efficiency achieved with each of these configurations
is shown in Figure 3 for intervals when the inlet S02 ranged from 1,050 to
1,250 ppm, the typical inlet level. Points shown on this plot represent
data taken during the normal course of operations in both periods when
inlet and outlet 862 levels and system pH's were simultaneously available
and confirmed. In most cases, each data point represents at least a few
'lours of operation at the condition shown. Where such a continuum of data
exists, outlet SC>2 levels and pH's have been averaged and rounded to the
learest 5 ppm and 0.05 pH units, respectively.
The data in Figure 3, which reflects the general operating experience
at Scholz, confirm the high SC>2 removal capability of sodium-based scrubber
systems operating in an equivalent concentration range of active sodium.
Achieving a given outlet S02 level (within the limit of the number of
contact stages in use) was essentially a. matter of adjusting the operating
pH of the scrubber system by changing the feed forward rate and/or pH of
the regenerated liquor.
Using both the venturi and absorber (with two trays) , outlet
levels below 50 ppm could be easily achieved at a venturi liquor pH above
5.2. This corresponds to better than 95% removal efficiency. Higher
removal efficiencies were attained with increasingly high pH levels;
however, there was little to be gained by operating at pH's much in excess
of 6.0. Above a pH of 6.0, outlet S02 levels dropped to 20 ppm or less
(>98% removal), taxing the accuracy and operating range of the outlet SC>2
monitor.
For the most part, when both the venturi and absorber were operated
together, the pH of the venturi bleed liquor was maintained between 4.8
and 5.9. Consequently, outlet S02 levels generally ranged between 15 and
100 ppm, and typically ran 30 to 70 ppm. In Period 2, for example, during
December and November the outlet S02 level ranged from 5 to 210 ppm and
averaged 45-50 ppm. During these two months there were only five occasions
when the outlet S02 exceeded 100 ppm. Three of these occurred while the
system was being operated with limestone rather than lime fed to the
regeneration reactor.
Operating with the venturi alone (liquor bypassing absorber trays)
a pH feed of 6.4-6.5 was required to achieve 95% removal (less than 50 ppm
outlet S02). However, better than 90% removal could still be quite easily
achieved at pH's on the order of 6.0. When the venturi alone was used in
Period 1, the bleed pH was generally maintained above 5.7 to keep outlet
S02 levels at or below about 100 ppm.
453
-------
500
400 -
300 .
Outlet SO2
(ppm)
Venturi Active
Ap.(in. H20) Na+,(M)
Operational Configuration
0.25-0.4
0.15-0.3
0.15-0.3
Venturi + 2 Trays
Venturi + 2 Trays
Venturi (No Feed to Trays)
1050-1250 ppm
5.5 6.0
Scrubber Bleed Liquor pH
Figure 3, SO2 Removal in the Scrubber System as a
Function of pH of the Scrubber Bleed Liquor
-------
2. Lime Utilization
In general, lime utilization was quite good, equalling that predicted
from the laboratory and pilot plant developmental work. Under normal
operating conditions, lime utilization ranged from 90% to 100% of the
available Ca(OH)2 in the hydrated lime (the hydrated lime ranged froa 87%
to 93% available Ca(OH)2). Based upon analysis of the solids produced,
lime utilization typically ran about 95% of the available Ca(OH)2. These
data are generally confirmed by overall system material balances.
During Period 1 it is not possible to accurately calculate the overall
Ca(OH)2/AS02 stoichiometry due to the frequent operational upsets, including
the occasional feeding of limestone. The best estimate for the first five
months of operation (after initial startup) is 0.95 to 1.00. This estimate
is based upon the integral averages of S02 inlet and outlet and lime feed
(adjusted for estimated unreacted limestone). During November and December
1975, when accurate material balances were closed, the Ca(OH)2/AS02
stoichiometry was 0.98. Sodium carbonate makeup represented an additional
alkali input to the system.
3. Oxidation and Sulfate Control
As would be expected, the major fraction of sulfite oxidation in the
system occurred in the scrubber circuit; and the single most important
variable influencing the oxidation rate was the oxygen concentration in
the flue gas. Estimated oxidation rates in the scrubber circuit ranged
from a low of about 150 ppm equivalent S02 (at an oxygen level of 5%) to
a high of about 400 ppm (at an oxygen level of 9%). By contrast, the
oxidation in the remainder of the process (the regeneration and dewaterlng
systems combined) was less than 25 ppm. Thus, oxidation in the scrubber
circuit accounted for more than 90% of the total system oxidation.
Figure 4 shows the estimated oxidation rates in the scrubber system
during both periods in equivalent ppm of S02 (design gas flow basis) as a
function of flue gas oxygen content. Included in this plot are data for both
scrubber system operational configurations.
It would be expected that oxidation rates would be somewhat lower
using the venturi alone than with the combined venturi and absorber, due
to the decreased gas/liquid contacting. However, such a decrease in
oxidation was not observed. This can be attributed to three factors:
first, the carryover of entrained liquor from the venturi to the absorber;
second, contacting of gas with liquor recirculated through the open spray
header; and finally, the flow of a small amount of regenerated liquor onto
the top tray through a leaky shutoff valve.
Considering all data, regardless of scrubber configuration, the data
scatter exhibited in Figure 4 amounts to a range of 80 ppm of oxidation at
a given flue gas oxygen level. This data scatter can be accounted for by
differences in operating temperature (120-135°F), liquor flow, and slight
differences in pressure drop in the tray tower, as well as sampling and
analytical errors.
455
-------
*».
tn
500
400
300
Sulfite
Oxidation
(ppm of SC>2 —
design gas flow basis)
200
100
~~ Venturi
Operational Configuration AP (in.
• Venturi + 2 Trays 5-7
• Venturi + 2 Trays 8-11
a Venturi (No Feed to Trays) 11 — 13
Operating
Period
2
1
1
5678
Oxygen Content of Flue Gas (% dry vol.)
Figure 4, Oxidation in the Scrubber System as a
Function of Flue Gas Oxygen Content
10
-------
Assuming the median value of oxidation to represent the average
oxidation experienced, then oxidation ranged from 175 ± 40 ppm at 5% oxygen
to 380 ± 40 ppm at 9% oxygen. For a typical S02 level of 1,200 ppm with
95% S02 removal and 90% gas load, these oxidation rates correspond to 17%
and 37% of the S0£ removed. Adding in the oxidation through the remainder
of the system, the total oxidation would be 20% of the S02 removed at a
5% oxygen level and 40% at 9% oxygen. For high sulfur coals (3-4% S),
these same overall oxidation rates would correspond to roughly 8% and 16%,
respectively.
It is worth noting that these oxidation rates represent operation with
three active contact stages (apparently even when the venturi was operated
alone). Normally in a low or medium sulfur coal application only one or
two stages should be required. This reduction in the number of stages
should substantially reduce oxidation rates.
At steady state the total sulfate formed must be removed from the
system by the precipitation of calcium sulfate and/or sodium sulfate losses
at the rate it is being formed. Calcium sulfate levels measured in the
product solids ranged from as low as a few percent of the total insoluble
calcium-sulfur salts during early startup to as high as 30% during periods
of high oxygen levels and low S02 (when the ratio of sodium sulfate to
active sodium in the system liquor was fairly high). The typical range of
calcium sulfate in the waste cake during Period 1 was 15% to 25% (mole basis)
of the total calcium-sulfur salts; and in Period 2 was 12% to 20% (mole basis)
These levels of calcium sulfate precipitation indicate that the system is
capable of keeping up with oxidation rates of up to 25-30% of the S02
removal in the concentrated mode of operation. In a situation where the
cake was thoroughly washed, the bulk of the sulfate losses would be as
calcium sulfate; sodium sulfate losses would represent less than 2-3% of
the S02 absorbed. The system liquor composition would simply self-adjust
the ratio of sodium sulfate to active sodium to precipitate the amount of
calcium sulfate required to keep up with oxidation.
The sulfate balances established during November and December 1975
are shown in Table 7. These months represent stable operating intervals
in which material balances were closed. During November, 20% of the S02
absorbed was oxidized to sodium sulfate in the system. Of this 20%, 13%
left the system as insoluble calcium sulfate. The remaining 7% left as
sodium sulfate. About half of the sodium sulfate was lost in the liquor
occluded with the cake (V>% sodium salts on a dry cake basis); and the
other half through uncollected pump seal leakage, the thickener leak, and
other unaccounted losses. The total sodium sulfate loss was calculated
from the soda ash makeup during this period.
In December, 23% of the 862 absorbed was oxidized to sulfate. About
15% left as insoluble calcium sulfate; about 3% as sodium sulfate occluded
with the cake (^3.5% sodium salts on a dry cake basis); and the rest through
system leaks and increase in liquor sulfate concentration.
457
-------
TABLE 7
SULFATE BALANCES PORING PERIOD 2
Average Inlet S02 (ppm)
Average S02 -Removal (Z)
Average Inlet 02 (vol Z, dry)
Total Sulfate Formation (Z AS02)
Sulfate Account (Z AS02):
In Cake - CaSOi*
- Na2S04
Na2SOit in System Liquor Inventory
Other
11/3-11/23/75
1,265
95.5
5.5
12/2-12/23/75^
1,135
96.5
6.0
Total
13.0
4.5
-0.5
4.0
21.0*
/j
15
3
2
3
23
.0
.0
.5
.0
.54
This NaaSOi* loss is calculated from the difference between total net
sodium makeup and sodium losses in the cake and represents sodium sulfate
losses in pump seal leaks, spills, and errors in estimates of soda ash
makeup.
458
-------
Had all system leaks been stopped or returned to the system and the
cake been washed to the same level of soluble solids, then the sulfate to
active sodium ratio in the liquor would have increased somewhat to allow
more calcium sulfate to precipitate, thereby re-establishing the equilibrium
between sulfate formation and losses.
4. Waste Cake Properties
The solids content of the filter cake produced during both periods of
operation typically ranged from 45% to 60% solids. In general, the solids
content decreased with increasing calcium sulfate content of the cake.
Process upsets (e.g., high 'levels of limestone in the lime feed) and
mechanical problems with the filter also tended to depress the level of
solids in the cake to the 45% range.
During intervals of normal process conditions with the filter operating
under standard operating conditions, the solids content of the cake usually
varied between 50% and 55% for calcium sulfate levels in the cake of
between 10% and 25% of the total calcium-sulfur salts. At this level of
solids, the cake generally had the appearance and consistency of a moist
powder.
During startup, when active sodium levels were fairly high (above 0.4M)
and sulfate concentrations were below 0.6M, the solids content of the cake
ranged from 55% to 70%, indicating the potential for producing even drier
cake under high sulfur coal conditions.
Wash efficiency tests conducted at Scholz during Period 1 indicated
that with two banks of wash sprays in series, the soluble solids content
of the cake could be reduced to 2-3% (dry cake basis) under controlled
filter conditions using a wash ratio (gals, wash water/gal. water occluded
in cake) of about 2.5. However, mechanical problems and process upsets
frequently prevented a reasonably continuous operation of the filter with
a thin enough cake to allow consistently high wash ratios and, therefore,
low solubles losses. The levels of solubles in the filter cake in Period 1,
therefore, ranged from 2% to 12% of the dry cake weight with an average of
5-8%. The 12% solubles level occurred during periods when the cake was not
washed.
Prior to the start of the second period of operation and during the early
part of Period 2, a third bank of wash sprays was added and the capacity
of'the existing sprays increased. This resulted in a doubling of the wash
capacity from 3 to about 7 gpm, which ensured that sufficient wash water
would be available during periods of high rates of cake withdrawal. In
Hovember, the level of soluble solids in the cake averaged 4.8% (dry cake
basis) with a wash ratio of 1.5-2.0. In December, the wash rate was
Increased to a wash ratio of 2.0-2.5 and the average soluble solids level
was decreased to 3.3% (a range of 1.5% to 6.0% dry cake basis). The
average solids content during both months was 51%.
459
-------
During Period 2, a limited amount of testing was performed with differ-
ent types of filter cloths. A number of different nylon and polypropylene
cloths In both monofilament and multifilament weaves were tried. The original
cloth, a polypropylene multifilament, was durable. However, it produced
poorer solids (45-50%) under the more adverse process conditions and tended
to blind, requiring cleaning of the cloth about two to three times a week.
The purpose of the testing was to eliminate or minimize the blinding while
maintaining the high solids content and durability. To date, the best
cloth has been a polypropylene monofilament with the same porosity as the
original multifilament. This cloth was used exclusively in December 1975.
The durability of the cloth is still being tested.
5. Sodium Makeup
The sodium makeup requirement in the operation of a dual alkali
system is determined simply by the rate of sodium loss, both controlled
(solubles in the filter cake) and uncontrolled (pump seal leaks, tank
spills, etc.). Under normal operating conditions the losses in the filter
cake should be the single most important sodium loss, and sodium makeup
should equal or only slightly exceed the quantity contained in the cake
(on a mole equivalent basis) in a tight, closed-loop operation.
Wash efficiency tests performed at Schoiz during Period 1 indicated
that 2-3% solubles losses in the cake cduld be achieved with a wash ratio
of 2.5. This solubles loss translates directly into an equivalent soda
ash makeup requirement of 2-3% of the S02 absorbed on a mole basis, which
represents a reasonable lower limit on soda ash makeup.
Soda ash makeup to the system has been consistently higher than this
level throughout the first year of operation. During the first half of
the year (Period 1) mechanical problems with the filter and Insufficient
washing limited the level to which soluble sodium salts could be washed
from the cake over an extended period to a minimum of about 5-8%. In
addition to the cake losses, there were also inadvertent spills as new
operators gained familiarity with the system as well as the normal leaks
from pump packing, piping, etc. Thus, soda ash makeup requirements
frequently exceeded 10% of the S02 absorbed on a mole basis.
The improvements in the filter operation and increased wash efficiency
in Period 2 substantially reduced sodium losses in the cake to 4.8%
solubles in November and 3.3% solubles in December. These sodium losses
corresponded to a soda ash makeup requirement equivalent to 5.8% and 3.6%
of the S02 absorbed (moles Na2C03 per 100 moles of S02 removed), respec-
tively. The net soda ash makeup (not including that used in increasing
the inventory of sodium In the system), though, was 10.5% in November and
7% in December. The higher makeup requirement in November is partially
due to contamination of the lime with 25 tons of limestone, increasing the
amount of waste solids and associated sodium losses.
460
-------
The apparent difference of 3-4% between the estimated soda ash feed
and sodium losses in the cake may be due to leakage from the system
(pump seals and the small thickener leak) and errors in estimates of the
soda ash feed.
Entrainment losses of sodium were very small. In particulate sampling
conducted during December, the total weight of sodium salts in the scrubbed
gas averaged 0.002 grains/scf. This represents a liquor loss through
entrainment of less than one gallon per hour. The soda ash makeup required
to replace this entrainment loss is less than 0.1 mole % of the S02 absorbed.
Efforts will continue throughout the remainder of the testing to
minimize soda ash makeup requirements by adequate washing of the filter
cake and further control of the leakage of liquor. The better cake
produced with high sulfur coal should allow reduction of solubles losses
in the filter cake to below 3%.
6. Process Reliability
Process reliability refers to the overall ease of process operation
including the resistance of the process chemistry to operational upsets,
the sensitivity of the process performance to small changes in the process
chemistry, and the potential for scaling in the process equipment. In
this regard it is to be differentiated from the mechanical/equipment-
related problems previously discussed.
In all respects the process reliability has been excellent. The
system has been successfully operated over a range of widely fluctuating
inlet S02 levels and oxygen concentrations in the flue gas, with little
or no change in the S02 removal efficiency, cake properties or lime
utilization. Throughout Period 1 with active sodium concentrations
between 0.15 and 0.30M, soluble calcium levels generally fluctuated
between 20 and 250 ppm and were usually below 150 ppm. In Period 2,
when active sodium levels were maintained above 0.3M, soluble calcium
concentrations averaged about 70 ppm and were consistently below 100 ppm.
In this range of calcium concentration, 20-250 ppm, there were no problems
with scale formation in the reactor (other than the mechanical problems
already discussed), the dewatering, or scrubber systems.
The low potential for scale formation and solids deposition in the scrubber
system was further demonstrated in Period 2 when the wash sprays beneath
the demister were turned off. For the three-month period from October
through December the demister was operated without wash water. There was
no increase in pressure drop across the demister nor the formation of a
scale or deposit of any kind.
The only period when difficulties were experienced due to changes in the
process conditions was in July when the inlet S02 levels dropped to 850-950
5pm and the flue gas oxygen level rose to 8.5-10.0%. The system was allowed
to drift into a dilute mode (active sodium levels decreased to less than
461
-------
0.15M). The very low soluble sulfite concentrations caused the soluble
calcium concentrations to rise to gypsum saturation levels and above,
resulting in the precipitation of gypsum and the deposition of gypsum
scale in the reactor vessels and reactor outlet piping. The scale was
removed during system shutdown. Had the active sodium been increased by
charging additional soda ash to the system, the gypsum scale would have
slowly redisaolved, as evidenced by the softening and eventual redissolution
gypsum scale that was not removed during the shutdown.
In addition to the fluctuating S<>2 levels and oxygen concentrations, the
system also was operated during a number of process upsets. These upsets
included the carryover of small amounts of fly ash (Period 1), the con-
tamination o€ the lime feed supply with limestone (Periods 1 and 2) , and
overfeeding of lime (Period 1). Only the limestone had any effect on the
pksjcess performance and this effect was temporary. During periods when
pure limestone was fed to the reactor system, the pH of the regenerated
liquor fell to below 7, causing a loss of S02 removal efficiency, from
above 90Z down to 80-85%. The solids content of the filter cake also
decreased slightly and sodium losses in the cake correspondingly increased.
Within a day or two after the limestone had passed through the system, all
of the effects were reversed and the system operation was back to normal.
462
-------
APPEMPIX
EQUIPMENT/INSTRUMENTATION STARTUP AND MAINTENANCE PROBLEMS
System
SCRUBBER
REACTOR
Problem
EQUIPMENT:
• Deterioration of refractory in reheat
burner chamber
• Slight corrosion and ash buildup on
fan during extended shutdown periods
• Separation of bond on rubber linings
in absorber and venturi bleed control
valves
• Hairline cracks and pinholes in
lining on absorber recycle tank
• Leaking of liquor through pump seals
and piping on pump suction lines
• Deterioration of fan thrust bearing
• Deterioration of stack lining due to
poor curing
INSTRUMENTATION:
• Water condensation in pressure taps
on level controllers
• Plugging of flow-through pH units on
recycle tanks
EQUIPMENT:
• Plugging of dry lime feed chute to
first reactor
• Buildup of solids in first reactor
due to location of dry lime feed
chute and poor agitation
Action
Replaced
Clean and rebalance fan after
extended shutdowns
Replace with SS valves under
warranty
Patched
Replaced packing and flange
gaskets
Replace
Patch/replace under warranty
Relocated lines
Relocate sampling lines
Installed vibrator
Design and install new first
reactor
Status
Corrected (Period 1)
Scheduled (1-2/76)
Corrected (Period 1)
Testing of new packing
(Period 2)
Scheduled (1-2/76)
Scheduled (1-2/76)
Corrected (Period 2)
Scheduled (1-2/76)
Corrected (Period 1)
Scheduled (1-2/76)
-------
EQUIPMENT/INSTRUMENTATION STARTUP AND MAINTENANCE PROBLEMS
Scrubber
REACTOR
(cont.)
Problem
FILTER/
THICKENER
• Broken agitator blade in second
reactor
• Broken shaft on reactor pump due to
piece of rubber lining from agitator
blade caught in impeller
• Failure of isolation valves on
reactor pumps
INSTRUMENTATION:
• Erosion and plugging of flow-through
pH unit
EQUIPMENT:
• Erosion of fiberglass scraper
• Erosion of plastic bridge valve due
to solids carried through cloth holes
• Loosening of cloth retaining ropes
out of caulking strips
• Loss of vacuum due to cracks in
internal fiberglass trunnion tubes
in filter
• Erosion/cracking of fiberglass rocker
arm on tub agitator
• Insufficient agitation in filter tub
to suspend sand-like particles and
grit
• Plugging of thickener underflow lines
• Deterioration of sections of lining
in thickener and thickener hold tank
Action
Status
Replaced agitator shaft and
impeller under warranty
Replace
Overhaul
Replaced with immersion type
probe with sonic cleaner
Replaced with SS under warranty
Instructed operators to shut
down filter and repair holes
immediately
Designed and installed new
temporary caulking strips and
reduced blower pressure
Patched cracks—new caulking
strip reduced stress on
internals when installing cloth
Reinforced with SS
Occasionally wash tub—(new
agitation system to be
installed)
Installed flexible lines and
new back-flushing provisions
Patch and replace lining under
warranty
Corrected (Period 2)
Scheduled (1-2/76)
Scheduled (1-2/76)
Corrected (Period 1)
Corrected
Corrected
Corrected
Corrected
(Period 1)
(Period 1)
(Period 2)
(Period 2)
Corrected (Period 1)
Scheduled (1-2/76)
Corrected (Period 1)
Scheduled (1-2/76)
-------
Scrubber
FILTER/
THICKENER
(cont.)
SODA ASH
GENERAL
EQUIPMENT/INSTRUMENTATION STARTUP AND MAINTENANCE PROBLEMS
Problem Action
INSTRUMENTATION:
• Poor level transmitter reliability in Replaced
thickener hold tank
EQUIPMENT:
• Clogging of dry feeder gate with lumps Replace feed control to allow
of soda ash
• Failure of circuitry in heat tracing
on piping
INSTRUMENTATION:
• Poor reliability of feed control
system for soda ash liquor
EQUIPMENT:
• Water freeze damage to pump seal
water rotameters
higher gate position
Replace
Replace
Replaced and adjusted operating
procedures
Status
Corrected (Period 2)
Scheduled (1-2/76)
Scheduled (1-2/76)
Scheduled (1-2/76)
-------
GLOSSARY
Active Sodium - Sodium associated with anions involved in SO- absorption
reactions and includes sulfite, bisulfite, hydroxide and carbon-
ate/bicarbonate. Total active sodium concentration is calculated
as follows :
2 x ([Na2S03J + [Na^]) + [NaHS03] + [NaOH]
Active Sodium Capacity - The equivalent amount of S02 which can be theo-
retically absorbed by the active sodium. Active sodium capacity
is defined by:
capacity - [Na2S(>3] + 2 x [Na^] + [NaOH] + [NaHC03J
Calcium Utilization - The percentage of the calcium in the lime or lime-
stone which is present in the solid product as a calcium-sulfur
salt. Calcium utilization is defined as:
Calcium Utilization - ™ls (CaS°3 + CaSV generated
mol Ca fed X 100Z
Concentrated Dual Alkali Modes - Modes of operation of the dual alkali
process in which the active sodium concentrations are greater
than 0.15M active sodium.
Dilute Dual Alkali Modes - Modes of operation of the dual alkali process
in which the active sodium concentration is less than or equal
to 0.15M active sodium.
Sulfate Formation - The oxidation of sulfite. The level of sulfate for-
mation relative to SO absorption is given by:
Sulfate Formation » mols S°3
mol S02 removed
466
100%
-------
Sulfate Precipitation - The formation of CaSO,» XH 0 in soluble solids.
The level of sulfate precipitation in the overall scheme is
given by the ratio of calcium sulfate to the total calcium-sulfur
salts produced:
Sulfate Precipitation = mols CaS°4
mol CaSO
X
TOS—Total Oxidizable Sulfur - Equivalent to the sum of all sulfite and
bisulfite species.
467
-------
REFERENCES
1. LaMantia, C.R., Lunt, R.R., and Shah, I.S., "Dual Alkali Process
for 862 Control," presented at Sixty-Sixth Annual Meeting, Amer-
ican Institute of Chemical Engineers (November 15, 1973), Paper
No. 25c.
2. LaMantia, C.R., Lunt, R.R., Oberholtzer, J.E., Field, E.L., and
Kaplan, N., "EPA-ADL Dual Alkali Program—Interim Results,"
presented at EPA Flue Gas Desulfurization Symposium, Atlanta
(November 4-7, 1974).
3. Lunt, R.R., Rush, R.E., Frank, I.E., LaMantia, C.R., "Startup and
Operation of the CEA/ADL Dual Alkali Process at Gulf Power/
Southern Services," presented at The American Institute of Chemical
Engineers, November 16-20, 1975.
4. Kaplan, N., "An Overview of Dual Alkali Processes for Flue Gas
Desulfurization," EPA Flue Gas Desulfurization Symposium, Atlanta
(November 1974).
468
-------
APPLICABLE CONVERSION FACTORS
ENGLISH TO METRIC UNITS
British
Metric
5/9 (°F-32)
1 ft
1 ft2 »
1 ft3
1 grain
1 in.
1 in.2
1 in.3
1 Ib (avoir.)
1 ton (long)
1 ton (short)
1 gal.
1 Btu
°C
0.3048 meter
0.0929 meters2
0.0283 meters3
0.0648 gram
2.54 centimeters
6.452 centimeters2
16.39 centimeters3
0.4536 kilogram
1.0160 metric tons
0.9072 metric tons
3.7853 liters
252 calories
469
-------
THE FMC CONCENTRATED DOUBLE-ALKALI PROCESS
L. Karl Legatski, Karl E. Johnson, and Lyon Y. Lee
FMC Corporation
799 Roosevelt Road
Glen Ellyn, Illinois
ABSTRACT
FMC Corporation first developed its Concentrated Double-Alkali
Process during the 1960's in response to internal needs. The system
has since been commercialized by the Company's Environmental Equipment
Division. The first full scale system was installed in one of the
Company's own plants in 1971. Since then, it has been tested on a
prototype scale (- 1 Mw) in ten different applications. A larger
(3 Mw) demonstration plant has been in continuous operation for over
one year, and the largest double-alkali system in the country was
brought on stream in October, 1975.
FMC's development strategy and operating experience for these
installations are discussed, and the economic and technical advantages
of the system are enumerated. The system has demonstrated outstanding
reliability and operability together with capital and operating costs
equal to or less than lime/limestone systems.
471
-------
INTRODUCTION
The FMC Concentrated Double-Alkali Process was initially
developed by the Company's Industrial Chemical Division (ICD)
in response to the need for a reliable flue gas desulfuri-
zation system for its own plants.
As early as 1956, ICD experimented with lime and lime-
stone scrubbers and encountered many of the now well-known
problems inherent in these processes. This early work led
to a program to develop a throwaway process that could meet
the following criteria:
1. High availability-
2. Easy to operate - within the capabilities
of existing operating personnel; and
3. Capital and operating cost competitive
with lime/limestone processes.
The FMC Concentrated Double-Alkali Process was the
result of this program. It was first utilized as a large-
scale prototype (about 30 Mw equivalent) in the Company's
Modesto Chemical Plant in 1971. Responsibility for commer-
cialization of the process for fossil fuel fired boilers
was subsequently transferred to the Environmental Equipment
Division. Since that time, the process has been tested and
refined on a pilot plant scale and in several different com-
mercial applications. A 3 Mw prototype has had 94% avail-
ability for the initial year of operation, and a 50 Mw
equivalent commercial system started up in October, 1975
has also had a high availability since that time.
In late 1975, FMC received a patent on the Concentrated
Double-Alkali Process covering what are believed to be the
most feasible operating conditions of the process. FMC
believes that its process has been sufficiently demonstrated
in large scale installations, and it is aggressively pursuing
large scale utility and industrial applications. The capital
and operating costs compare very favorably to lime and lime-
stone systems, and the reliability and ease of operation are
demonstrated facts.
472
-------
PROCESS DESCRIPTION
Figure 1 is a process schematic for the FMC Concentrated
Double-Alkali Process. Essentially, the process entails a
sodium based scrubbing loop in which sulfur dioxide is collected
according to the reaction:
Na2SO3 + SO2 + H20 •> 2NaHSO3
and subsequently precipitated with calcium in a separate loop
according to the reaction:
2NaHSO3 + Ca(OH)2 -»• CaSO3 + Na2S03 + 2H2O
The principal advantages of the process in comparison to
other types of throwaway processes are:
1. No scale in the scrubber;
2. Ease of control of the process•
3. Superior chemical utilization;
4. Improved solid waste properties• and
5. Ability to control sulfur dioxide
and flyash concurrently.
One of the concerns in the double-alkali process
is the oxidation of the sulfite ion in the scrubbing solution
to sulfate. In contrast to the sulfite ion, the sulfate ion
is not readily precipitated with the addition of calcium hydroxide
and the sulfate ion therefore must be purged from the process at
a rate equivalent to its net formation rate in the process. The
loss of sodium compounds from the process is in the form of en-
trained liquor associated with the calcium sulfite and flyash
insolubles. This loss is a problem in that alkaline sodium
chemicals are more expensive than alkaline calcium chemicals
and the soluble salts leaving the process may pose a potential
water pollution problem. However, in comparison to lime and lime-
stone processes, this is not a disadvantage in that the high
magnesium limes that have been successful in lime scrub-
bing systems often produce a waste for disposal with a solubles
content greater than or equal to a double-alkali process. Addi-
tionally, use of FMC's Concentrated Double-Alkali Process re-
sults in virtually stoichiometric utilization of chemicals
and produces a mechanically stable filter cake of low permea-
bility. The filter cake need not be fixed, which results in
total operating costs substantially less than those of conven-
tional lime/limestone systems.
473
-------
FLUE
GAS
BY-PASS
TO EXHAUST
STACK
CQ(OH)2-
DISC
CONTACTOR
SCRUBBER
LIME REACTOR | C«aS03
SOLID TO DISPOSAL
SCRUBBING
SO-f NcSO
REGENERATION
NOZS034-2HZ0
Fiqure 1. Schematic of FMC's Concentrated Double-Alkali Process.
-------
In developing its double-alkali process, FMC was guided
by the following objectives:
1. Develop the scrubbing solution composition
and scrubber design that would minimize the
net formation rate of soluble sulfate ion;
2. Utilize a scrubbing solution and regeneration
solution chemistry that would be easy to
control within the constraints of existing
process control technology and be responsive
to large and rapid changes in boiler operating
conditions ;
3. Develop the scrubbing chemistry that would
minimize liquid flow rates and liquid holdup
requirements in order to minimize the capital
and operating costs of the facility ; and
4. Develop a regeneration system that would yield
a solid waste product physically and chemically
acceptable for landfill disposal while minimiz-
ing the size and complexity of the equipment
required for liquid-solid separation.
To accomplish these objectives, FMC has studied the
chemical mechanisms involved in the process, the unit
operations, the process controls and the materials of
construction. The rate of soluble sulfate formation has
been characterized as a function of the solution composition,
flue gas composition, and scrubber design. Scrubber design
has been optimized in terms of scrubber efficiency, pres-
sure drop, and mist elimination capabilities. The chemistry
of the lime regeneration step has been studied to maximize
chemical utilization and minimize equipment size require-
ments. Thickener settling rates have been studied under a
wide variety of operating conditions and vacuum filter per-
formance has been investigated in terms of alternative
filter design features and filter cake washing character-
istics. The filter cake product has been evaluated chemically
and physically on a laboratory and prototype scale for accepta-
bility as a landfill. Materials of construction have been
and continue to be routinely evaluated for acceptability
under the various chemical and physical conditions experi-
enced in the process. A proprietary control system for
the -process has been designed and demonstrated by virtually
thousands of hours of trouble free operation.
The following sections briefly describe the state of
FMC's development efforts in the various areas of process
technology.
475
-------
Scrubber Performance
The original type of scrubber utilized by FMC for sulfur
dioxide control was its proprietary Dual Throat Venturi
Scrubber. It was felt that a scrubber capable of simul-
taneous flyash and sulfur dioxide collection in a single
contacting stage would offer significant advantages over
the scrubbers utilized with lime and limestone processes
which required flyash collection in a venturi followed by
sulfur dioxide collection in an absorber. All of FMC's
early pilot plant experience was with a venturi type
scrubber, and the simultaneous collection capability was
adequately demonstrated.
For stoker fired boilers, the flyash concentration could
typically be reduced to 0.1 Ibs/MM Btu simultaneously with
sulfur dioxide collection efficiencies of 90% at scrubber
pressure drops of approximately 10" w.g. However, the ven-
turi type scrubber cannot be successfully operated at pres-
sure drops significantly less than 10" w.g. without the risk
of deterioration in collection efficiency. Furthermore, the
tendency of utilities in recent years to utili2e high efficiency
electrostatic precipitators to comply with particulate regu-
lations has decreased the potential market for simultaneous
flyash and sulfur dioxide removal systems.
In 1973, FMC developed and tested its proprietary Disc
Contactor Scrubber. This scrubber, which is essentially a
baffle type scrubber, is designed to allow sulfur dioxide
collection efficiencies in excess of 90% at a relatively low
pressure drop without the use of spray nozzles. The original
prototype was an 800 cfm unit that was tested on a boiler
operated by Colorado Public Service Company in Boulder,
Colorado. Subsequently, a 2000 acfm prototype unit was built
and operated in conjunction with the trailer mounted pilot
plant at St. Joe Minerals Corporation. The unit demonstrated
an average collection efficiency of 94.9% operating on a pul-
verized coal boiler burning 2% sulfur coal.
While particulate removal was not a requirement of the
scrubber, particulate removal efficiencies were also deter-
mined. When operating a scrubbing system with a scrubbing
solution containing substantial quantities of dissolved or
suspended solids, the possibility of the particulate emissions
being adversely affected by entrainment of the scrubbing liquor
is a potential concern. However, in a dissolved salt system,
it is possible to utilize mesh type mist eliminators which
provide excellent entrainment removal. This was demonstrated
by tests where the scrubber inlet grain loadings, already
below 0.01 gr/scf, were further reduced by the scrubber.
476
-------
Furthermore, during a test where the precipitator was
malfunctioning and was out of compliance with the 0.02 gr/scf
guarantee, the Disc Contactor Scrubber was able to reduce the
grain loading to within the specified level.
An additional test was conducted to determine the actual
extent of entrainment from the scrubbing solution because the
high sodium sulfate concentrations utilized in a concentrated
double-alkali process may adversely affect total sulfate emis-
sions. The results of these tests indicated that the entrain-
ment of sodium sulfate was negligible.
With either type of scrubber, the Dual Throat Scrubber or
the Disc Contactor Scrubber, FMC utilizes a scrubbing solutior
with a pH of approximately 6.5. If the pH is above 7, carbon
dioxide absorption becomes significant and can lead to calcium
carbonate scaling. A scrubbing solution pH below 6 is like-
wise avoided because the vapor pressure of sulfur dioxide
increases dramatically for concentrated systems and can lead
to equilibrium limited scrubbing conditions where outlet con-
centrations below 200 ppm are desired. This fact is demon-
strated in Figure 2 in which the sulfur dioxide vapor pressure
is plotted as a function of pH for a solution temperature of
130°F, a typical saturation temperature for a boiler flue gas
stream. The vapor pressure of sulfur dioxide over dilute solu-
tions is not significant, as can be seen from the line in the
figure calculated from the Henry's Constant for sulfur dioxide
and water and the first and second ionization constants for
sulfurous acid for a solution with a total oxidizable sulfur
(TOS) level of 0.03 gram moles/liter and an ionic strength of
0.3 gram moles/liter. However, for concentrated solutions,
the sulfur dioxide vapor pressure at a pH of 6 is approximately
100 ppm and increases rapidly with decreasing pH.
For these reasons, FMC prefers to operate the scrubbing
side of its process in the pH range of 6 to 7, and preferably
at a set point of 6.5 where the sulfite-bisulfite system is
highly buffered. The highly buffered system is preferred be-
cause the scrubbing solution is able to adapt very well to
rapid changes in flue gas inlet conditions. This is in con-
trast to scrubbing systems utilizing a pH of 6 or less which
can be very sensitive to rapid changes in flue gas inlet con-
centrations.
Sulfate Formation Rate
Superficially, the concept of the double-alkali process
is quite straightforward. Sulfur dioxide is absorbed with a
sodium sulfite solution, and the sodium sulfite solution is
regenerated by precipitating the sulfur dioxide as calcium
sulfite. The principal difficulty with the process is that
sodium sulfite reacts very rapidly with oxygen to form sodium
477
-------
-------
sulfate, and the sodium sulfate cannot be readily regenerated
into an active sodium form by reaction with lime. One of the
principal areas of diversity in double-alkali process tech-
nology is the subject of sulfate formation and sulfate removal
from the process.
FMC has found the effects of flue gas and solution compo-
sition on the rate of sulfate formation to be significant. A
major part of FMC's development work has been to evaluate the
effects of these parameters on the net formation rate of sol-
uble sulfate. FMC first observed that, everything else being
equal, increased ionic strength, principally in the form of
increased sulfate concentration, reduced the net soluble sul-
fate formation rate. The results obtained during the pilot
plant program indicated that a high ionic strength double-
alkali process can be expected to produce a net soluble sul-
fate formation rate on stoker fired boilers operating with
high excess air and flyash present equivalent to less than
10%, but probably not less than 5%, of the sulfur dioxide
collected when operating with 2.5% or greater sulfur coal.
The effect of oxygen concentration on net soluble sulfate
formation rate was also investigated. The results clearly
indicated that with a pulverized coal boiler operating with
4 to 7% oxygen in the flue gas stream, the equivalent of
approximately 30 to 60 ppm of sulfur dioxide will leave the
process as sodium sulfate. While FMC has never been able to
isolate any catalytic effect on oxidation caused by the presence
of flyash, sulfate formation rates have generally been higher
for stoker fired boilers with flyash present than with pulver-
ized coal fired boilers with no flyash present. However, no
data are available for low oxygen concentrations on stoker
fired boilers and very little exists for high oxygen concen-
trations on PC fired boilers.
Inlet sulfur dioxide concentration has not been found to
influence soluble sulfate formation. The sulfur dioxide con-
centrations encountered in pilot operations ranged from 400 to
8000 ppm. From these experiments, no correlation between net
Soluble sulfate formation rate and inlet sulfur dioxide con-
centration could be found.
time Reactor Performance
For the regeneration system, FMC utilizes the concept of
regenerating the sodium bisulfite by reaction with lime according
to the following reaction:
Ca(OH)2 + 2NaHS03 -> Na2S03 +
479
-------
To accomplish this reaction, FMC utilizes a low residence
time continuous stirred tank reactor. The lime reactor is
controlled at a pH of about 8.5, which is effectively the
titrametric endpoint for sodium bisulfite.
It is possible to precipitate additional sulfur values
at elevated pH's according to the following reaction:
Ca (OH) 2 -f Na2SO3 -»• CaSO3 + 2NaOH
However, it has been FMC's experience that it is difficult
to prevent excess lime addition at pH's in excess of about 10.
The responsiveness of the pH control system is excellent in
the presence of bisulfite ion, but is relatively poor at the
higher pH's existing in the lime-sulfite region. If a high
pH set point is used, changes in scrubber operating conditions
which lead to changes in the flow rate to regeneration can
lead to the addition of excess lime due to the poor respon-
siveness of the pH control system. Excess lime leads to poor
chemical utilizations and also results in poor filter cake
quality. Furthermore, substantial precipitation of sulfite
leads to increased calcium solubility in the system and poten-
tial super saturation. The only advantage of operating the re-
generation system at a high pH is the potential for minimizing
the flow rate to regeneration by precipitation of additional
sulfur values from the sodium sulfite solution.
FMC prefers to utilize a relatively high bisulfite con-
centration {typically 0.3 M) in the scrubbing solution which
provides a flow rate to regeneration equal to or less than
that obtainable in more dilute systems that attempt both to
regenerate the bisulfite and precipitate additional sulfite
by reaction with the sodium sulfite. The virtue of operating
the lime reactor at the bisulfite endpoint is further demon-
strated by the consistently high lime utilizations achieved
by FMC. Independent analyses of the filter cake conducted by
Firestone Tire and Rubber Company, General Motors Corporation,
and St. Joe Minerals Corporation all indicate free lime con-
centrations in the filter cake of less than 1% on a dry basis.
The concentrated double-alkali process has the additional
virtue of "built-in" softening due to the high dissolved sul-
fite concentrations always present in the scrubbing solution.
This results in a dissolved calcium concentration an order of
magnitude below its saturation level. Consequently, calcium
sulfate scaling in the FMC process is virtually impossible.
In designing its lime reactor, FMC determined that for
operation in the pH range of 8 to 10, residence times of only
a few minutes were sufficient with reactor temperatures in
excess of 120°F. At temperatures below 120°F (which may exist
480
-------
for applications such as acid plants) steam heat has been
found to be an effective means of elevating the temperature
to a level that will support the lime reaction.
Liquid-Solid Separation
FMC's Environmental Equipment Division is a leading sup-
plier of thickening and clarification equipment for water
treatment applications. The design of the thickening and
clarification equipment for concentrated double-alkali pro-
cesses is dependent upon many variables including the quality
of the lime, the composition of the scrubbing solution, and
the reactor temperature. FMC has experimented with rotary
vacuum filters produced by several different manufacturers
and prefers a knife discharge with an air puff to a belt
filter because it requires less maintenance and produces a
drier cake.
Vacuum filter performance results obtained from one in-
stallation recently are typical. These show an average per-
cent solids in the filter cake of 61.3% for a two month period.
The filter cake washing efficiency was approximately 70%. This
was accomplished with a wash rate of from 2 to 3 displacement
washes, reducing the soluble content of the filter cake on a
wet basis to from 2 to 3%.
Landfill Studies
As a regular part of an ongoing development program, FMC
has evaluated the filter cake product produced from this pro-
cess for acceptability as landfill. These tests have been
underway in laboratory and prototype demonstrations since 1973
under the direction of Dr. Raymond Krizek, Professor of Soil
Mechanics, at Northwestern University. The tests originally
concentrated on the physical properties of the material to
determine its handleability, transportability, and bearing
strength as a landfill material. It was found in the labor-
atory tests, and later confirmed in pilot plant demonstrations,
that the material was amenable to transport on conveyor belts,
could be loaded into trucks and subsequently moved, and showed
a surprisingly good bearing strength.
Further tests were conducted to determine the leachability
of the material. These tests showed that, while the soluble
salts contained in the filter cake could be removed in shake
tests, the material itself was highly impermeable, and proper
Management of a landfill could prevent displacement of the
soluble salts by rain water.
481
-------
A prototype landfill was constructed in conjunction
with the Firestone Demonstration Plant. The landfill was
designed according to good engineering practice, with appro-
priate test wells and reference wells plus provision for
periodic core samples. These tests have been in progress
over one year and, at this time, no contribution of soluble
material to the ground water has been observed. Furthermore,
the physical characteristics of the landfill have been extremely
encouraging. The solid material is distributed in the pit area
by a front end loader which maneuvers in the area without sig-
nificant difficulty. Photographs of the filter cake and the
landfill area are shown in Figures 3 and 4 and serve to demon-
strate the favorable overall characteristics of the material.
FMC recognizes that some states may require fixation
and/or the use of pond linings for disposal of the filter
cake material, but will endeavor to prove that the material
is physically and chemically acceptable for disposal in a
non-chemical landfill without the use of chemical additives.
There is no evidence to date that contradicts the above sup-
position, and the additional tests being conducted should con-
firm or refute FMC's position. In the worst case, a liner
may be necessary, and initial cost estimates indicate that an
appropriate lining material could be provided for less than
$1.00/ton of disposal material. The filter cake produced by
FMC's process should never have to be fixed to provide mechan-
ical stability.
Water Balance
The FMC process is designed to dispose of all of the
water that enters the process as either evaporation or en-
trained moisture on the filter cake. The degree of difficulty
in achieving this goal depends on the sulfur dioxide concen-
tration and the temperature of the flue gas stream. Higher
sulfur dioxide concentrations require more lime slaking water
and more filter wash water. Most of the water that leaves the
process is in the form of evaporation, and the evaporation
loss is directly related to the inlet flue gas temperature.
Thus, a flue gas at 400°F containing 1500 ppm sulfur dioxide
does not pose a water balance problem, while a flue gas stream
at 280°F containing 3000 ppm sulfur dioxide will have a very
tight water balance. For example, in the latter case, two
displacement washes on the filter cake may consume 50% of the
total makeup water requirement and lime slaking water may con-
sume an additional 30 to 40%. The washing and lime slaking
requirements, therefore, may consume virtually the entire
water makeup requirement of the process.
The only uses of water in the process other than the filter
washing and lime slaking water requirements are the pump seal
water, soda ash solution makeup, and mist eliminator washing
requirements.
482
-------
V ,«** ^«* i •»
Figure 3. Landfill area at Firestone (note dozer tracks
in right foreground).
<»". ~fr~
>l >**
<-' ^ ,J~ v1*
m * m^ ^^ •
*<•*« -Vfc'_-4c.
v
Figure 4. Climbing a recently dumped three-foot deep pile of
waste material demonstrates physical stability.
483
-------
OPERATING EXPERIENCE
FMC's operating experience can be summarized in five
different applications or areas of experience:
-1. The original installation at FMC's
Modesto Chemical Plant;
2. Subsequent pilot plant work-
3. The Firestone Demonstration Plant;
4. The Caterpillar/Mossville installation; and
5. FMC's Green River Plant installation.
The most important facts concerning these projects are dis-
cussed below.
Modesto Chemical Plant
In 1956, the Modesto Chemical Plant of FMC's Industrial
Chemical Division began work on an evaluation of the various
possible processes for sulfur dioxide removal from flue gases.
The Modesto Chemical Plant produces barium and strontium chemi-
cals, such as the oxides, chlorides, carbonates and nitrates,
by reducing barium and strontium sulfate in a reduction kiln
to produce the respective oxides, which are reacted with the
appropriate acids or soda ash in other parts of the plant.
The reduction operation produces relatively high sulfur dioxide
concentrations fluctuating in the range of 4000 to 8000 ppm.
The Modesto staff, with assistance from the Industrial
Chemical Division's Princeton Research Center, initially tested
lime and limestone scrubbers, but encountered many of the now
classical problems with these systems. They next piloted a
sodium based scrubbing system using soda ash, since soda ash
was manufactured by FMC and already in use in the plant in
other areas. The soda ash system proved to be very effective,
but disposal of the sodium sulfite and sulfate products posed
a problem. FMC was ultimately able to negotiate a contract
with a nearby paper mill to sell the sodium sulfite and sulfate
for makeup to their paper process. Unfortunately, just as
484
-------
detailed design of this system was completed, the paper mill
cancelled the arrangement. Thus, the Modesto staff was faced
with finding a method of recovering the sodium values from
the sodium based scrubbing system while producing an accept-
able solid waste for disposal. In response to this necessity,
they developed the concentrated double-alkali process. This
process was, in fundamental areas, the same as the process
sold commercially today by FMC. The double-alkali process was
successfully demonstrated by the Modesto staff on a pilot scale,
and subsequently installed and operated on a full scale in 1971.
Figure 5 is a photograph of the Modesto operation.
This unit, which processes a flue gas stream of 30,000 acfiu
is equivalent to approximately 30 Mw on the chemical regener-
ation side of the process. While the unit is not completely
representative of the design FMC would utilize today, the system
has demonstrated an extremely high degree of reliability in an
application experiencing severe fluctuations in the process in-
put conditions. Since startup, the process has been available
in excess of 95% of the time, the majority of the non-available
time accounted for by kiln down time rather than scrubber down
time.
Pilot Plants
Shortly after the startup of the Modesto facility, the
Air Pollution Control Operation began an evaluation of the
potential for commercialization of sodium scrubbing and double-
alkali processes. The processes were found to offer significant
operating advantages in coal-fired applications in comparison
to the competing lime and limestone processes, and the capital
and operating costs appeared to be very competitive. Further,
it seemed that if sulfur dioxide could be adequately collected
in a venturi type of scrubber using a concentrated sodium based
scrubbing system, the combining of flyash and sulfur dioxide
collection into one operation would offer significant capital
cost advantages.
In the Spring of 1972, a pilot test program was performed
at Modesto using a venturi scrubber on a slip stream from the
kiln gas system. This was the first in a long series of pilot
plant experiments with the double-alkali process. This pro-
gram demonstrated that a venturi type scrubber, such as FMC's
Dual Throat Scrubber, could successfully collect sulfur dioxide
in excess of 90% efficiency using the concentrated double-alkali
process. Inlet sulfur dioxide concentrations of from 4000 to
8000 ppm were routinely reduced to the 100-200 ppm level.
Based on this encouraging experience, a pilot plant of the
regeneration part of the process was built to operate in con-
junction with the pilot scale venturi scrubber. This unit was
installed on a slip stream from a stoker fired boiler at FMC's
485
-------
Figure 5.
Concentrated Double-Alkali System installed
at FMC's Modesto Chemical Plant.
.
-------
Industrial Chemical Plant in South Charleston, West Virginia
in late Spring, 1972.
The test at South Charleston lasted for approximately
six months and consisted largely of experiments related to
scrubber performance. The boiler was very old and seldom
fired at full load, so the oxygen concentration was very
high, typically 12 mole percent. This fact, combined with
the relatively low inlet sulfur dioxide concentrations, which
fluctuated from 400 to 1400 ppm, led to oxidation rates that
represented a significant fraction of the sulfur dioxide col-
lected. Thus, the majority of the effort at South Charleston
was devoted to determining the effects of solution composition
and scrubber operating conditions on the sulfur dioxide col-
lection efficiency and the rate of soluble sulfate formation.
At the conclusion of the test program, it was decided that
the process was technically feasible for application to coal
fired boilers. The Air Pollution Control Operation was incor-
porated into the Environmental Equipment Division, and the de-
cision was made to pursue the sulfur dioxide control market in
the industrial boiler field. The pilot plant was installed in
a 35 ft. van to provide efficient transportation to the facili-
ties of potential customers. In the Spring of 1973, a contract
was obtained from Caterpillar Tractor Co. for a two month
demonstration of the process on a slip stream from a stoker
fired boiler at the heating plant of Caterpillar's Mossville
Engine Plant.
The pilot plant was operated on an 8 hour per day basis
for two months by two FMC Engineers, and the results were com-
pletely satisfactory to Caterpillar and consistent with the
prior pilot plant experience by FMC. As a result of this suc-
cessful operation, Caterpillar Tractor Co. awarded FMC a
contract in May 1973 for turnkey installation of the sulfur
dioxide control system at the Mossville Engine Plant.
The pilot plant program was ultimately extended to ten
different applications in which the process was applied for
periods from a few weeks to several months. The range of
the input variables experienced by the system was substantial:
SO2 concentrations - 400 to 8000 ppm
0- concentrations - 5 to 12%
Temperature - 120 to 500°F
Particulate - Nil to 3.0 gr/sdcf
These varied inputs allowed FMC to develop a knowledge of the
phenomena that govern sulfite oxidation and sodium consumption
that is unsurpassed in the industry.
487
-------
Firestone Demonstration Plant
In January of 1975, FMC completed the installation of
a three Mw equivalent demonstration plant (Figures 6 and 7)
for The Firestone Tire & Rubber Company in Pottstown, Pennsyl-
vania. This installation, utilizing a 10,000 acfm Dual Throat
Venturi Scrubber, is operating on a slip stream from a pulver-
ized coal boiler that is presently burning 2% sulfur oil to
meet particulate emission regulations in the State of Pennsyl-
vania. While this unit is not of commercial scale and is not
officially a commercial operation, the experience at Firestone
demonstrates four very important points about the process.
First, the installation is of major significance because
virtually all of the process data and relationships that were
developed in the pilot plant program have been confirmed in
this installation, which is four or five times as large as
the trailer mounted prototype unit.
The second key feature of the project is the proven oper-
ability of the system. Two weeks after startup, the operation
of the unit was turned over to three boiler house operators
(one per shift) who have been responsible for the operation
since that time. FMC provides one engineer on the day shift
to perform chemical analyses and experiments related to the
performance of the various unit operations, but he has vir-
tually no responsibility for the routine operation of the
system. Performance has been so good and the system has re-
quired so little operator attention, that a remote alarm
system is now being installed to allow the operators to work
their normal shift at the boiler control panel and inspect
the demonstration plant area only once every two hours or
whenever a common alarm sounds at the boiler panel indicating
an upset in the process.
Third, the landfill study program discussed in a pre-
ceding section has demonstrated the handleability and imper-
meability of the filter cake waste material.
Finally, the availability of the operating system has
been outstanding. For the entire first year of operation,
including most of the start-up period, the overall availability
has been 93.5%. This is particularly impressive in view of
the lack of any spares in the system and the fact that the
operation was often unattended on the night shift.
The various segments of the system that contributed to
the downtime are itemized in Table 1. The single longest
outage was due to blockage of the thickener underflow pipe
by a plastic beaker that was unintentionally dropped into the
thickener. The second longest outage was caused by improper
maintenance of the recirculation pump seal, which led to
separation of the pump lining from the pump housing. The
488
-------
Figure 6. FMC Dual-Throat Scrubber and Cyclone in
Firestone Demonstration Plant.
489
-------
Figure 7.
Lime storage tank and dumpster at Firestone
Demonstration Plant.
490
-------
Table 1.
AVAILABILITY OF FIRESTONE DEMONSTRATION PLANT
Downtime
Item
1. Thickener Pluggage
2. Pump s
3. Cake Conveyor
4. Fan
5. Lime Feeder
6. Spray Nozzles
7. Instrumentation
8. Control Valves
% of Total Downtime
20.7
16.1
16.0
15.4
13.4
12.4
4.1
1.9
100. 0
% of Operating Year
1.35
1.04
1.04
1.00
0.87
0.80
0.27
0_.13_
6.50
Total Annual Availability = 93. 5i
Note: The Firestone Demonstration Plant has no installed
spares.
491
-------
filter cake conveyor problems generally related to the
fact that an inappropriate conveyor was used, due to the
unavailability of high quality equipment for the small
size required. The fan outage was due to overheated
bearings, which were in turn caused by shaft misalign-
ment. The remaining outages were similarly due to problems
unrelated to the process, and all of them could have been
eliminated or substantially reduced by a normal commercial
sparing philosophy.
Caterpillar Mossville Installation
FMC's largest double-alkali installation to date is the
system for Caterpillar Tractor Co. at its Mossville Engine
Plant outside of Peoria, Illinois. The unit is the largest
double-alkali system in the country and is designed to col-
lect sulfur dioxide and flyash from four stoker fired boilers
totaling 460,000 Ibs/hr of generating capacity. The operation
of this unit commenced in October, 1975. Figures 8 and 9 il-
lustrate portions of the system.
During startup and initial testing, the Caterpillar Moss-
ville unit has demonstrated a high availability. At the present
time, two of the boilers are in operation, and the system
has not yet been tested at full load. Furthermore, one of
these boilers is new and has been operated only intermittently
because it is still being debugged. However, there has been
only one occasion since startup in which a boiler has been
shut down due to a problem with the FGD system, and the scrubbers
have not been modified or maintained during the boiler outages.
FMC expects the high availability achieved during startup to
continue or improve during sustained boiler operations.
One of the key features of the Mossville installation is
the simultaneous collection of flyash and sulfur dioxide.
There are no mechanical collectors on these boilers, and the
full flyash load is removed by the scrubbers. This fact, com-
bined with the high excess air inherent in the stoker boilers,
creates conditions that are far more technically difficult for
a double-alkali process than a low excess air pulverized coal
boiler with a high efficiency precipitator attached. At Moss-
ville, most of the problems to date have been related to wear
with the control valves, filter cloth, and conveyor belts.
At this time, FMC feels that it has arrived at satisfactory
solutions to all of these problems. In the long run, the con-
centrated double-alkali process offers the potential for re-
liable, simultaneous collection of sulfur dioxide and flyash
in utility and industrial applications. This would result in
substantial savings in capital and operating costs, and FMC
feels that this will be one of the major advantages of double-
alkali systems in the future.
492
-------
Figure 8. Dual-Throat Scrubber (80,000 ACFM) being installed
on 150,000 lb./hr. boiler at Caterpillar Tractor Co
493
-------
Figure 9. Control panel at Caterpillar Mossville installation,
494
-------
Based on FMC's performance to date on the Mossville
project, Caterpillar Tractor Co. has awarded FMC contracts
for the design of complete S02 and particulate control
systems for its East Peoria and Mapleton plants. Each
system will have a heat input equal to approximately 100 Mw.
The engineering design for these units has been completed
and ultimate startup of the two systems is expected to be
in 1977 and 1978, respectively.
FMC Green River
This Spring, FMC will start up its largest sodium
scrubbing system which consists of two 330,000 acfm Disc
Contactor Scrubbers on the new pulverized coal fired
boilers now being installed at FMC's Green River Soda Ash
Plant in Green River, Wyoming. The boilers are two
625,000 Ibs/hr units followed by high efficiency electro-
static precipitators. The scrubbers are designed to remove
90% of the sulfur dioxide at a pressure drop of 5.5" w.g.,
and are equipped with a bypass reheat arrangement.
This represents the third different type of reheat
arrangement used by FMC in commercial installations. The
Caterpillar Mossville facility, due to spatial limitations,
uses Dual Throat Scrubbers followed by steam tube reheaters
operating on an ambient air stream injected into the scrub-
ber discharge, followed in turn by an induced draft fan.
The East Peoria and Mapleton installations for Caterpillar
utilize forced draft fans on the scrubbers, followed by
provision for steam tube air injection reheat. The use of
steam tube air injection reheat was dictated by Caterpillar
to reduce its dependency on fuels other than coal. Due to
the low sulfur dioxide collection efficiency requirement
at Green River, the scrubbers are designed to utilize by-
pass flue gas for reheat whenever possible and oil fired
reheat when higher sulfur coal is burned or the boiler is
at full load.
495
-------
CAPITAL AND OPERATING COSTS
It is difficult to generalize on the capital and
operating costs for a flue gas desulfurization system,
because of the differences in design bases and input
assumptions that can occur. Due to significant operating
advantages of the double-alkali system, many people have
arrived at the conclusion that it is inherently more
expensive than lime/limestone. Closer examination of both
systems indicates that this conclusion is not justified.
In capital cost, it is superficially obvious that
double-alkali and lime/limestone processes must be com-
parable. In fact, the comparison in Table 2 indicates
that double-alkali may be less expensive. Virtually
every component in the double-alkali process has an
equal or larger counterpart in a lime/limestone process.
The only potential disadvantage of double-alkali is that it
requires filters, but many lime/limestone systems now
use filters.
In operating costs, the concentrated double-alkali
system is generally less expensive in all areas. Chemical
consumption is essentially equal to lime systems; limestone
may be less depending on location.
Power requirements are less due to lower pressure
drops and liquid rates. Water requirements are slightly
less. Reheat is often less because the high efficiency
of double-alkali makes bypass reheat possible. Maintenance
and operating labor are obviously less in the highly buffered,
soluble salt, concentrated double-alkali system. Concentrated
double-alkali systems do not scale, and scrubbing with a
solution instead of a slurry results in substantially
less wear.
Finally, solid waste disposal is substantially less ex-
pensive. There is less material to begin with because of the
superior chemical utilization and the low moisture content of
the filter cake. Chemical fixation is not required for
mechanical stability.
A typical example of estimated operating costs for a
300 Mw Midwestern utility burning 3.13% sulfur coal is
given in Table 3.
496
-------
Table 2. QUALITATIVE CAPITAL COST COMPARISON OF
DOUBLE-ALKALI AND LIME/LIMESTONE SYSTEMS*
Equipment Item
1. Scrubbers
2. Pans
3. Pumps
4. Thickener
5. Filter
6. Tankage
7. Materials in
8. Materials out
Lime/Limestone
12 to 15" w.g.
typical
L/G = 30 to 40
typical
Less or Equal
Less storage for
1ime s tone, but
substantially more
material used
FMC Concentrated
Double-Alkali
Equal or smaller
(less mass transfer
units)
Smaller
(6" w.g. or less)
Smaller
(L/G = 10 gal/mcf)
Equal
Less (low residence
time)
About same as lime
Substantially less
*No flyash removal included.
497
-------
Table 3. ESTIMATED OPERATING COSTS FOR THE FMC CONCENTRATED
DOUBLE-ALKALI SULFUR DIOXIDE REMOVAL SYSTEM1
Chemicals
Lime (92% CaO)
Soda Ash
Consumption
0.047 Ib/KWH
0.0021 Ib/KWH
Unit Cost C/KWH
$33/ton 0.0776
$64/ton 0.0067
Utilities
Power
Water
Fuel Oil2
0.0137 KWH/KWH
0.078 gal/KWH
0.0006 gal/KWH
1.75C/KWH 0.0240
40C/Mgal 0.0031
35C/gal 0.0210
Operating
Labor &
Supplies
1 man/shift +
supplies
$30,000/man 0.0067
year +
$50,000
supplies
Maintenance
3% of Plant
0.03x$21 MM 0.0240
Solid Waste
Disposal
0.204 Ib/KWH
$4.00/ton
0.0408
TOTAL OPERATING COST (excluding capital charges)
0.2039
300 Mw unit operating at full load with 3.13% sulfur coal;
90% SO2 collection; system designed for 10% over heat
input using the maximum sulfur, minimum Btu coal expected.
250°F of total reheat: 25°F with oil, and 25°F with bypass
flue gas.
498
-------
CONCLUSIONS
In conclusion, there are substantial advantages to
the FMC process in comparison to other systems. In com-
parison to other double-alkali systems, FMC offers the
following:
1. FMC has more cumulative hours of operating
experience in pilot plant, prototype, and
commercial systems on more different appli-
cations than anyone else. Due to the
breadth of its experience, FMC is in a
position to specify performance stan-
dards and accept significant financial
obligations as a consequence of potential
nonperformance.
2. The sulfur dioxide collection capabilities
of FMC's process and equipment are demon-
strated facts. With sulfur dioxide control
systems there is a potential risk that the
scrubbing liquor entrainment will adversely
affect the particulate and/or primary sul-
fate emissions. FMC has demonstrated that
its Disc Contactor Scrubber does not adversely
affect fine particulate or primary sulfate
air emissions.
3. In terms of process considerations, FMC has
demonstrated essentially 100% utilization of
makeup chemicals (soda ash and lime) through
thousands of hours of prototype and commer-
cial operation. Scale prevention is inherent
in the FMC process due to sulfite softening
of the scrubbing liquor in the regeneration
portion of the process. At no time in either
its commercial or prototype experience has
FMC experienced a scaling problem in its
scrubber or elsewhere in its process. FMC
has an excellent knowledge of the causes of
sulfate formation and has demonstrated the
capability to predict and inhibit the
net sulfate formation rate it will experience
by controlling the process operating conditions.
4. The principal objective of all of FMC's develop-
ment work has been to develop a process that
takes advantage of the inherent reliability of a
sodium scrubbing process while striving to keep
the process simple enough to compete in capital
499
-------
and operating costs with lime and limestone
systems. The inherent reliability of the
process has been demonstrated in the Modesto
Plant of FMC and the Firestone Demonstration
Plant. The overall simplicity of the process
is apparent from the process description con-
tained in this section.
5. FMC's Double-Alkali Process is proprietary and
is patented (U.S. Patent No. 3,911,084). FMC
believes the operating conditions specified in
the patent are optimal for operating a concen-
trated double-alkali process.
In comparison to lime/limestone processes, FMC believes it
offers the following significant advantages:
1. Ability to achieve higher SO2 removal efficiencies.
This often allows the use of bypass reheat result-
ing in a substantial savings.
2. Sodium scrubbing processes are much more toler-
ant of process upsets and variations in gas
flow and composition.
3. The reduced liquid-to-gas ratio (usually 10 gallons
per 1000 acfm) results in smaller pumps and tanks.
4. The double-alkali process produces a drier and
more mechanically stable filter cake.
5. The sodium scrubbing processes produce a lower
outlet particulate loading due to no slurry in
the scrubber. This allows the use of a mesh type
mist eliminator without plugging.
6. The S02 collection efficiency is variable over
wide limits by simply adjusting the chemistry.
This can be significant in cases where a high
SO- collection efficiency is not desirable or
necessary.
7. The scrubbing solution is highly buffered so cor-
rosion is reduced. This also results in a more
operable and less sensitive system.
8. Calcium scaling does not occur. FMC has never
shut down a scrubber because of scaling in an
SO2 application.
500
-------
9. One of the most significant advantages of
sodium scrubbing is the ability to collect
SO and flyash concurrently. This can re-
sult in substantial capital cost savings.
10. FMC's installations are being operated by
boiler house personnel. All installations
have had in excess of 90% availability, and
the average has been about 95%. The demon-
strated reliability and operability of FMC's
installations is unsurpassed in the industry,
to our knowledge.
FMC continues to develop and refine its basic process
through an intensive research and development program in
the following areas:
1. Improved scrubber design to reduce
capital and operating costs;
2. Improved thickener design to reduce
space requirements;
3. Improved filter cake washing to further
reduce sodium losses; and
4. Continued landfill evaluation program to
demonstrate the acceptability of disposal
without further treatment.
501
-------
ACKNOWLEDGEMENTS
FMC Corporation gratefully acknowledges the assistance
of Messrs. Roman Zaharchuk and Gary W. Wamsley of The
Firestone Tire & Rubber Company and Mr. David C. Dietrich
of Caterpillar Tractor Co., all of whom made material con-
tributions to the success of the projects discussed in this
paper.
502
-------
OPERATING EXPERIENCE WITH THE ZURN DOUBLE ALKALI FLUE GAS
DESULFURIZATION PROCESS
P. M. Lewis
Zurn Industries, Inc.
Air Systems Division
Birmingham, Alabama
ABSTRACT
The first Zurn Double Alkali Flue Gas Desulfurization went on-line
in September, 1974, at the Joliet Plant of Caterpillar Tractor Company,
Joliet, Illinois. The system is scrubbing the flue gas from two coal
fired boilers rated at 100,000 Ib./hr, and 80,000 Ib./hr. steam. These
boilers produce steam for building heat and are in service about seven
months a year. The system utilizes a dilute double alkali process with
lime addition for regeneration of the spent scrubbing solution, and
sodium carbonate addition for scrubber feed water softening. The calcium
based sludge generated by the process is dewatered for disposal in a
remote landfill.
This paper summarizes the process, and discusses operating experiences
to date. Process reliability, equipment performance, and solutions to
problems are discussed.
503
-------
OPERATING EXPERIENCE WITH THE ZURN DOUBLE ALKALI FLUE GAS DESULFURIZATION
PROCESS.
I. INTRODUCTION
Due to the uncertain future availability and cost of natural gas and fuel
oil, Caterpillar Tractor Co. decided in 1970 to rely on coal for steam
generation in two boilers at the Caterpillar Tractor Co. plant in Joliet,
Illinois.
Basically, there are two methods for complying with sulfur oxide emission
standards while burning coal. These are, removing the sulfur oxides from
the flue gas before it is emitted to the atmosphere, and the use of low
sulfur coal. The demand for low sulfur coal was increasing rapidly, and
its future price and availability prospects were not favorable. The use
of high sulfur coal and flue gas desulfurization appeared to be the most
economical alternative that would provide an adequate fuel supply and ensure
compliance with current regulatory requirements. In particular, after
technical and economic evaluations, they selected the Zurn Double Alkali
Flue Gas Desulfurization System.
It is the purpose of this paper to describe this system and to relate the
operating experience at the Joliet, Illinois installation to date.
II. SYSTEM DESCRIPTION
The Zurn Double Alkali Process uses a dilute solution of sodium hydroxide
to absorb the sulfur oxides from flue gas. Lime is used to regenerate the
spent scrubbing solution to form sodium hydroxide and insoluble sulfur
bearing calcium compounds which can be removed by mechanical means.
This double alkali process can be viewed as having three functionally
different process stages. As shown in Figure 1, these process functions
are gas scrubbing, chemical regeneration, and solids removal.
GAS SCRUBBING
The function of the gas scrubbing section is to intimately mix the flue
gas containing sulfur dioxide with a dilute sodium hydroxide solution.
This reacts with the sulfur dioxide gas and converts it to water soluble,
sulfur containing compounds (sodium sulfite, sodium bisulfite, and sodium
sulfate). The scrubber also removes a high percentage of any fly ash or
other particulate which may be contained in the flue gas.
504
-------
The reactions that take place in the scrubbing section of the system
are shown below:
2 NaOH + S02 -t>- Na2S03+H20 (1)
Na2S03+S02-m20 -£>- 2NaHS03 (2)
Na2S03+l/202 -O- Na2S04 (3)
Sodium hydroxide and sulfur dioxide react to form sodium sulfite (reaction
1). The sodium sulfite then reacts with sulfur dioxide and water to form
sodium bisulfite (reaction 2). Some of active sulfite is oxidized to
sodium sulfate by excess oxygen in the combustion gases (reaction 3).
After being mixed with the flue gas, the solution is withdrawn from the
scrubber to an external mix tank for chemical regeneration of the spent
scrubbing solution.
REGENERATION
The purpose of the chemical regeneration section is the conversion of the
spent scrubbing solution to a usable form, and the separation of this
solution from the sulfur containing solid compounds. In the mix tank,
calcium hydroxide, from previously slaked quick lime, is added to the
spent scrubber solution. The calcium hydroxide reacts with the soluble
sodium salts to form insoluble calcium salts (calcium sulfite and calcium
sulfate (or gypsum). This reaction is effected in two agitated process
tanks and the resultant mixture is then pumped to a thickener where the
calcium salts precipitate and are removed as a slurry. The regenerated
sodium hydroxide solution overflows to a clarifier. The reactions in
the regeneration section are:
NaHS03 + Ca(OH)2 -£=- NaOH + CaS03 y + H20 (4)
Na2S03 + Ca(OH)2 -t>~ 2NaOH + CaS03 y (5)
Na2S04 + Ca(OH)2->- 2NaOH + CaS04 Y (6)
Sodium carbonate is added to the clarifier to further reduce the concen-
tration of calcium by the precipitation of calcium carbonate (limestone).
This step is referred to as "softening". The further precipitation of
calcium in this step reduces the likelihood of calcium salts precipitating
in the scrubbing section. The softening reaction is as follows:
Na2C03 + Ca(OH)2 ->• 2NaOH + CaC03 1 (7)
After the solution has passed thru the clarifier, it has been regenerated
and is available for reuse in the scrubbing section to remove more sulfur
dioxide The insoluble calcium salts have been removed in separate
process 'equipment specifically designed for solids precipitation and
concentrat ion .
505
-------
LIQUID-SOLID SEPARATION
The underflow slurry from the thickener and clarifier is pumped to the
solids removal section of the process for further dewatering. The
function of the solids removal section is to recover the usable scrubbing
solution contained in the thickened underflow slurry and to convert the
slurry solids into a form which is easily handled and is acceptable for
disposal. This is accomplished with rotary drum vacuum filters which
convert the slurry into a cake containing approximately 50 to 607o by
weight solids, with the remainder being water. These filters also have
provisions for "washing" the cake to remove as much of the soluble sodium
compounds as possible before disposal. The filter cake is composed of
calcium sulfite, calcium sulfate (gypsum), calcium carbonate (limestone),
and the particulate removed from the flue gas by the scrubber. *
The filter cake discharges to an automated conveyor system which alter-
nately fills waste product containers for off-site disposal as landfill.
The filtrate separated from the cake by the vacuum filters is recycled
back to the thickener to be reused in the system. Thus, the process
operates in a "closed loop", which means there is no process solution
discharged from the system other than that contained in the filter cake.
III. PERFORMANCE CHARACTERISTICS
DESIGN BASIS
The Zurn Industries, Inc. system installed at Caterpillar Tractor Co.'s
Joliet, Illinois plant cleans the flue gas from the two stoker fired
boilers which produce steam for building heat and process needs. These
boilers are rated at 80,000 and 100,000 Ibs. of steam per hour and each
is equipped with a mechanical dust collector. Total coal firing rate to
these boilers is 9.65 tons per hour and the process is designed to allow
the boilers to operate within the allowable emission rate of 1.8 Ibs. of
S02 per million BTUs heat input. System design parameters are summarized
below:
No. of Boilers 2
Firing Method Stoker
Total Steam Capacity 180,000 Lbs./Hr.
Total Gas Volume 103,500 Cu. Ft./Min.
Coal Sulfur Content 4% As Received
S02 From Boilers 6.74 Lbs./mmBTUs
Allowable S02 Emission 1.8 Lbs./mmBTUs
The total process design, engineering, construction, and start-up of this
system were accomplished by Zurn Industries, Inc. Design, construction,
and start-up were completed on, or ahead of schedule. The chronology of
system design and installation was as follows:
506
-------
1. Feasibility investigations 15 Months
2. System design 12 Months
3. Equipment installation started February 1, 1974
4. Component check-out started August 12, 1974
5. First scrubber put in service September 25, 1974
6. Second scrubber put in service October 2, 1974
EQUIPMENT - PROCESS PERFORMANCE
The scrubbers, breeching, stack, and scrubber effluent piping are con-
structed of 316 stainless steel. The wet surfaces of the vacuum filters
are 304 stainless steel. All other process piping and process tanks are
carbon steel. The lime mix tanks are lined with a Urecal polyurethane
coating.
System start-up and preliminary testing was accomplished from September
25 through December 16, 1974. The major reasons for downtime during this
period were minor mechanical equipment problems typical of system start-
ups, and the accumulation of unslaked lime and grit in the lime mix tanks.
This accumulation was due to poor slaking characteristics of the delivered
lime, high insolubles content, and some air-slaking of the lime in the
storage silo. Accumulation rates were quite dramatic. Approximately 10
feet of sand, gravel, and unslaked lime were deposited in the primary
lime mix tank in one ten day period. This problem was solved by the
installation of a lime slaker which improved lime slaking and had pro-
visions for separating sand and gravel from the lime before it entered the
system. The slaker installation was accomplished during the scheduled
Christmas 1974 shutdown.
Design revisions were completed and the system was put back on line January
23, 1975, and operated continuously until June 1, 1975. Process performance
and reliability during this period was quite good. One of the operation
highlights observed during the prolonged period of operation was good system
control over a wide range of system loads. The scrubber pressure drop and
pH were very stable, easily controlled, and were not adversely affected by
rapid boiler load changes. No significant accumulation of solids was
observed in the scrubbers which would impede their efficiency or operation.
No measurable corrosion or erosion was observed in any of the process
vessels, scrubbers, or piping.
Various mechanical problems with process auxiliaries were encountered during
this period and some of these are illustrated below.
Chemical Feed System
The lime and soda ash are metered to the process from outside silos by
screw conveyors. Two problems were encountered here. One was the relia-
bility of components in the electronic control system for the conveyor
drives and the other was due to formation of lumps inside the silos caused
by moisture from rainwater leakage and condensation on the inside tank walls,
507
-------
Both of these problems cause inaccurate feed rates which in turn caused
undesirable deviations in system chemistry. The conveyor drive control
systems were made operational by replacing faulty electronic components.
Inaccurate feed rates due to lumps of material forming inside the silos
was a continuing problem throughout this period of operation. A dry air
purge system was installed on the lime and sodium carbonate storage silos
to prevent material degradation and lump formation due to moisture. This
has effectively prevented lump formation in the storage silos, and dry
material flow from the silos has been reliable.
pH Probe Reliability
Accurate pH indication is essential to maintaining good system control,
and preventing scaling and corrosion. The pH probes as initially installed
were quite susceptible to coating and fouling which resulted in erroneous
readings. Several methods were investigated to alleviate this problem and
reliable pH monitoring was finally achieved by the installation of parallel
probes with fresh water back flush and ultrasonic cleaning.
Sulfur Dioxide Monitoring
Stationary probes are installed in the scrubber inlet duct and in the
stack to provide continuous monitoring of flue gas for SC>2 concentration.
The S02 analyzer was not operational during initial system check-out due
to shipping damage. After damaged components were repaired, the analyzer
was calibrated and put in service. Sample probe pluggages. which rendered
the S(>2 analyzer inoperable, became a recurring problem. Two system
revisions were made to improve probe reliability. The pressure of the
blow-back air to the probes was increased to improve probe cleaning and
shields were installed around the probes. These revisions did minimize
plugging but after a period of continuous operation the shields deteriorat-
ed. This problem was not resolved during the first operating season before
summer shutdown. Operation of the analyzer system has been good during the
1975-1976 heating season after revisions were made to the probe assembly
and purge cycle by the instrument supplier.
Water Balance
The process is designed for closed loop operation, and therefore should
have zero scrubbing solution discharge during steady state operating con-
ditions. Extended periods of zero discharge were not achieved during the
system start-up period for the following reasons: (1) Excessive water
was added to the system thru process pump seals and vacuum filter belt
wash sprays. (2) The boilers were often operated at low loads, which
resulted in low water evaporation rates in the scrubber.
The loss of scrubbing solution was alleviated by installating rotometers
on process pump seal water supply lines and by using clarified scrubber
feed solution for vacuum filter belt wash. The system requirement for
fresh water was significantly reduced by the installation of a seal water
508
-------
recirculation system on the vacuum pumps. Closed-loop operation has
been satisfactorily demonstrated since these modifications have been
made.
Freeze Protection
The system as originally installed, did not provide adequate protection
from freezing temperatures. Frozen process liquor lines and control air
lines contributed to varying degrees of abnormal operation. Installation
of additional insulation and higher quality heat tracing has provided
reliable freeze protection.
Feed Water Pump
Feed water pump impeller damage due to cavitation was experienced as a
result of insufficient suction head in the surge tank which serves as a
reservoir for the scrubber feed water pumps. The surge tank height was
increased to provide additional system surge capacity and to provide
greater suction head to the scrubber feed water pumps. This modification
has been very effective in preventing cavitation.
Vacuum Filter Belts
The system has two vacuum filters,, each capable of handling the total system
solids load. The filter belt originally used was woven of multifilament
polypropylene. Belt life was poor due to the accumulation of solids in
the interstices of the cloth. It was thought that the belt wash sprays
were ineffective, so higher pressure water was piped to the sprays. To
date, this modification appears to have prolonged belt life, but other
belt materials and cloth weaves are being tried in an effort to extend
belt life.
The system was put in service on October 24, 1975, for the 1975-1976 heating
season. After approximately two days of operation, the sludge pump impellers
began plugging with pieces of compacted sludge that had the consistency
of stiff clay. The entire process was idle from June until October, and
apparently the solids that could not be removed by filtration during shut-
down had compacted on the clarifier walls and rake mechanism. When the
process was restarted, this material fell from the metal surfaces, entered
the sludge pump, and greatly reduced pump capacity. This situation had
never been encountered prior to the first summer shutdown, and a revised
shutdown procedure will prevent a problem of this nature in the future.
The lime slaker required considerable maintenance and attention during
the first operating season, so hydrated lime was substituted for pebble
lime to decrease the maintenance requirement on this equipment. Operation
with hydrated lime has been good from a maintenance standpoint, but the
hydrated lime has different transport characteristics from the pebble lime,
and control of lime feed rate has been poor. Modifications to the volumetric
509
-------
lime feeder have greatly improved lime feed control.
Miscellaneous Hardware Performance
Most of the auxiliary equipment and controls have given good service and
have performed adequately. Operating experience and familiarity with
the system hardware has improved component reliability through a better
understanding of equipment capability and improved preventive maintenance.
Rubber lined diaphragm valves have given excellent service in the process,
and have replaced ball valves in some areas.
The radiation type density meters intended to indicate clarifier underflow
slurry density were removed due to drift. Slurry density is now monitored
by the operator.
Addition of a polyelectrolyte flocculant has yielded the expected favorable
results. Clarifie-r overflow clarity is excellent, slurry filterability
has improved, and preliminary results indicate that the cake solids content
has been increased from about 541 to around 62% by weight.
The automated sludge conveyor which alternately fills waste containers,
has performed well with little operator attention. Periodic cleaning is
required to remove sludge from idlers during extremely cold weather.
The Dustraxtor scrubbers, Figure 2, have been reliable and stable even
during rapid boiler load fluctuations. The entrainment type scrubber has
shown no tendency to scale or plug under any operating conditions. Some
soft solids accumulation has been found at the liquid surface in the
lower portion of the scrubber, but this does not increase pressure drop or
decrease operating efficiency. The solids appear to dissolve away at a
rate which prevents any significant accumulation. The demisting section
of the scrubbers is bare-metal clean with no solids accumulation whatever.
The difficulties with the lime slaking and lime feed control systems caused
some scale formation in the lime mix tanks during the first operating season
Improved process control has eliminated this problem.
IV. PROCESS EFFICIENCY AMD ECONOMICS
PROCESS EFFICIENCY
A comprehensive system performance test program has just been completed,
and only preliminary results of these tests are available. Preliminary
performance tests conducted by using the installed S02 analyzer and
independent wet chemistry analyses indicated that S02 removal capability
of the process is in excess of 85%. Particulate removal capabilities are
yet to be determined.
510
-------
PROCESS ECONOMICS
A prime reason for selecting the double alkali system was that it afforded
an economic advantage over the use of low sulfur coal or conversion to
low sulfur oil. Precise cost determinations are difficult due to
variable costs of raw materials and associated services, but the best
current estimates of system operating costs are shown below:
Basis : 4% S coal
75% S02 removal required Per Ton
of Coal
Chemicals (lime and sodium carbonate) $2.56
Water 0-02
Electricity 0.71
Waste 0.64
Subtotal $3.93
Labor (operation and maintenance) $2.12
Maintenance 0.38
Depreciation
Estimated Total Cost Per Ton Coal
V. SUMMARY
The Zurn Double Alkali System at the Joliet plant of Caterpillar Tractor Co.
was an advance ment in the technology of flue gas desulfurization and involved
certain unknowns. However, the system was made operational on schedule and
has encountered few problems, none of which haven't been effectively solved.
The process is currently in its second operating season and reliability
continues to steadily improve. The system has completed approximately 5600
hours of operation, including 2500 hours of satisfactory continuous operation
prior to planned summer shutdown.
System development and optimization are continuing to maximize the potential
of the process. In the near future, the remaining minor mechanical problem
will be resolved and more data will be available on system performance.
511
-------
ZURN DOUBLE ALKALI
DESULFURIZATION PROCESS
RAW MATERIALS - UTILITIES USAGE
LIME
SOM ASH
WATER
ELECTRICITY
WASTE DISPOSAL
PER HOUR
1040 LB.
285 LB.
36 GAL.
6.85 KWH
5200 LB.
PER TON COAL
108 LB.
29.5 LB.
3.7 GAL.
0.71 KWH
539 LB.
PER 1000 LB. STEAM
5.78 LB.
1.58 LB.
0.2 GAL.
0.038 KWH
28.9 LB.
BASIS: 180,000 #/HR. STEAM GENERATED
9.65 T/HR. COAL FIRED
4% SULFUR COAL
75% S02 REMOVAL REQUIRED
512
-------
FRESH WATER
BOILER
NO. 2
DRY CHEMICAL
TRANSFER
_^|VACUUM
rc°-RECEIVE
VACUUM
PUMP
VACUUM
FILTER
SPARE
EXISTING
COLLECTOR
VACUUM
RECEIVER
SPARE
BOILER
NO. 3
DUSTRAXTOR
VACUUM
PUMP
SPARE
CONTROL TANK
EXISTING
COLLECTOR
WASTE PRODUCT
(10) PICK-UP
CONTAINERS
FIGURE ZURN DOUBLE ALKALI SYSTEM
-------
I
FIGURE 2
ZURN DUSTRAXTOR SCRUBBER
514
-------
KUREHA FLUE GAS DESULFURIZATION
"SODIUM ACETATE-GYPSUM PROCESS"
Shigeru Saito, Toraijiro Morita, and Shigeyuki Suzuki
Kureha Chemical Industry Co., Ltd.
1-8 Nihonbashi Horidome-Cho, Chuo-Ku,
Tokyo, Japan
ABSTRACT
This paper deals with a new sodium acetate-gypsum desulfurization
process. This process is a complete closed loop with a S0? removal
efficiency of more than 99% using limestone or lime as regenerants
and can be applied to flue gases from oil fired as well as coal fired
boilers.
Furthermore, a new simultaneous removal of S02 and N0_ is also
described.
515
-------
KUREHA FLUE GAS DESULFURIZATION
" SODIUM ACETATE-GYPSUM PROCESS "
1. INTRODUCTION
Several years ago, Kureha first developed a sodium
sulfite process as a single alkali process to produce sodium
sulfite for the pulp and paper industries. Then, based on
this development, Kureha has further developed a sodium
sulfite-gypsum double alkali process in cooperation with
Kawasaki Heavy Industries, Ltd. The first full scale plant
of this process has been in operation at Shin Sendai power
station of Tohoku Electric Power Co., Inc. since April 1974.
Recently, it has also been proven by two full scale plants
of Shikoku Electric Power Co., Inc. having around 740,000
st.cu.ft./min. of flow rate.
Through this experience, we have learned that this
double alkali process is superior to other existing processes
with respect to SOp removal efficiency, operability and
economy.
However, the processes that we have developed, have been
considered mainly for flue gases from oil fired power plants
using sodium hydroxide or calcium carbonate as feed materials
to produce sodium sulfite or gypsum as by-products.
On the other hand, there are strong demands for the
removal of S02 and particulate in the flue gas from coal fired
boilers as well as flue gas from sinter plants of steel mills.
516
-------
Also, lime can, in some cases, be utilized more advan-
tageously than can lime stone. There is a strong incentive
to develop a process to meet the above requirements.
Key differences in the nature of flue gases from coal
fired boilers lie in the higher oxygen content, particulates
and chlorine when compared to flue gases from oil fired
boilers.
For the development of a new process, these factors
have been taken into our serious consideration and acetic
acid has been adopted as the best absorption liquor to
satisfy the basic requirement for a scrubber absorbent:
1. It should give a weak alkaline solution with large
buffering capacity, when its aqueous solution is
neutralized by strong base.
2. It should be a weaker acid than sulfurous acid, but
stronger acid than carbonic acid.
3. It should have high solubility in water, and the vapor
pressure should be as low as possible.
4. Solubility of its calcium salt in water should be large,
5. It should be easily available in the market and the
price should be reasonable.
517
-------
Acetic acid thus selected, has been utilized success-
fully as an efficient absorbent in the process development
through stepwise tests from 60 st.cu.ft./min. bench scale
test to 5,000 st.cu.ft./min. pilot scale test for the flue
gas from oil fired power stations.
Additionally, we have carefully investigated the
feasibility for flue gas from a coal fired boiler using
60 st.cu.ft./min. test facility.
Through a series of tests, we could overcome problems
that had been encountered in the cleaning of flue gases from
coal fired boilers. We have also obtained the necessary
information for the scale-up of this process.
We are confident that this process is superior in
operability and economy to the other known processes. Also,
in the present situation in Japan, hydrated lime can be used
advantageously as a regenerant for this process over lime
stone from overall process economy.
We are now in the planning stage of several full scale
plants in Japan to demonstrate the superiority of this
sodium acetate gypsum process using either limestone or
lime for flue gases from oil fired as well as coal fired
boilers.
We wish to report the results of our process develop-
ment of the sodium acetate-gypsum process. 1) 2)
518
-------
2. PROCESS DESCRIPTION
2-1. Lime Stone Method
As shown in Figure I, this process is composed of three
basic sections; absorption, oxidation and gypsum formation.
This process has been tested at 5,000 st.cu.ft./min. pilot
test facility for the flue gas from our oil fired power
station from March 1975 through August 1975 with continuous
and successful operation.
2-1-1. Absorption Process As illustrated in Figure 1,
the absorption tower consists of two sections - S02 absorp-
tion and acetic acid elimination section - , which are linked
in series in one absorption tower. The lower section is
composed of two absorption chambers and the higher acetic
acid elimination section consists of three chambers, each
being equipped with perforated plates inside of the chambers.
In the S02 absorption section, S02 is removed by circulating
a sodium acetate solution through the two stages.
The following reaction takes place:
2CH,COONa + S02 + H20 -» Na2S03 + 2CH3COOH (l)
A part of the acetic acid thus farmed volatilizes in
the scrubbed flue gas, and must be eliminated in the next step
In this step, fresh lime stone slurry is added to the
top chamber, and flows down count ercurrently from chamber to
chamber to the bottom to accomplish the complete removal of
acetic acid vapor. Also, this step serves to complete the
519
-------
elimination of remaining SC>2 in "the flue gas from the
absorption section, so that clean exhaust gas is released
into the atmosphere from the top of the scrubber.
For the smooth operation of the scrubber, prevention
of scaling is essential. Thus, the following four measures
have been taken against scaling prevention.
1. The structure of the scrubber has been designed in
such a way as ;there can be formed no dry parts inside
the scrubber walls.
2. Perforated plates have been used successfully in
preventing scaling, because vigorous mixing of flue
gas and absorption liquor through the plates have
resulted in both preventing scaling to and removing
scaling from the walls or the plates.
3. Seed crystals have been added successfully to prevent
scaling by the possible precipitation of calcium
sulfate in the absorbent. Especially, this precipita-
tion tends to take place,when the solubility of calcium
sulfate in the absorption liquor becomes lower at the
outlet of the absorption section than at the inlet.
4. When the concentration of sodium sulfite is relatively
high in the liquor, this reacts with calcium sulfate
in the liquor to form insoluble calcium sulfite, which
may cause a scaling problem. Accordingly, it is
necessary to keep the concentration of sodium sulfite in
the absorption liquor lower than 2$ as can be understood
from Figure 2. This can be successfully accomplished in
520
-------
our process by diluting the scrubber absorbent with
a part of oxidized liquor from the oxidizer.
By the application of these measures, we have achieved
the prevention of scaling in the scrubber and at the same
time a high efficiency of S0? removal and acetic acid
elimination for a variable loadings of gas flow rate as
summarized in Table 1.
2-1-2. Oxidation Process Sodium sulfite formed in the
S02 absorber is oxidized to sodium sulfate in the oxidation
tower which is provided with perforated plates to facilitate
fine dispersion of air bubbles and to promote oxidation.
Sulfite oxidation to sulfate takes place as follows:
S0 + 1/2 0 -> NaS0 (3)
After oxidation, the liquor is sent to the gypsum
formation section, where gypsum is produced by the addition
of lime stone slurry.
2-1-3. Gypsum Production Process In this step, gypsum
is produced from the oxidized liquor containing mainly
sodium sulfate and acetic acid by the addition of lime
stone slurry. In this gypsum formation, calcium carbonate
reacts very rapidly with sodium sulfate in the presence of
acetic acid, whereas the reaction of calcium carbonate
with sodium sulfate is very slow in the absence of acetic
acid.
In this process, calcium carbonate is assumed to react
at first with acetic acid present in the liquor to form
521
-------
calcium acetate, which reacts further with sodium sulfate
to form calcium sulfate and sodium acetate by a double
decomposition reaction. The regenerated sodium acetate is
recirculated to the scrubber after separation from gypsum
by filtration.
The reaction can be expressed as in (4) and (5):
2CH5COOH -> (CH3COO)2Ca + R^Q + C02 (4)
(CH3COO)2Ca + Na2S04 -» CaS04 + 2CH3COONa (5)
Gypsum produced in this step has such chemical composi-
tion as represented in Table 2.
This gypsum is high in quality and is utilized in wall
board making and in Portland cement production.
Thus, use of acetic acid as an absorbent has made it
possible to remove S02 efficiently from flue gas, producing
high quality of gypsum as a by-product.
Also, rapid reaction of sodium sulfate with calcium
carbonate has made this process further possible to apply
it to flue gases with high oxygen content such as flue gases
from coal fired boilers, which will be mentioned in the latter
section in detail.
2-2. Acetate-Lime Process
In order to make a comparison between lime stone and
lime, tests have been conducted at 60 st.cu.ft./min. bench
test facility to compare the date with those obtained by
lime stone process.
522
-------
In principle, acetate-lime (hydrated lime) process has
the same structure with the acetate-limestone process as
given in Figure 3. The absorbent containing sodium acetate
is in a similar way transferred from the absorption tower
through the oxidation process to the gypsum formation step,
where lime is added instead of lime stone to produce gypsum.
This reaction takes place more rapidly than with lime stone
even in a very dilute lime suspension. Also , the mother
liquor obtained after separation from gypsum crystals shows
a higher PH value than in the case of lime stone process.
These two features make it possible to operate the acetate-
lime process with advantage over lime stone process especially
in terms of removal efficiency of SC>2 and acetic acid removal
in the scrubber.
At the same time , concentration of the absorption liquor
can be reduced considerably to minimize the possible loss of
acetic acid by the oxidation.
Through these experiments, we have found that the
scrubber unit for S02 can be simplified from two absorption
sections to single section and as to acetic acid vapor
removal from three to one section, respectively. Also,
make-up of acetic acid has been reduced considerably.
In the simplified scrubber, the elimination section
of acetic acid by aqueous lime solution played only a
supplementary role, because most of acetic acid vapor in the
flue gas had been eliminated in the first S02 absorption
section by the absorption liquor of high pH value (7.0 - 7.2).
Figure 4 supports the above fact, in which relationships
between pH values of the liquor and concentrations of acetic
523
-------
acid in the gas phase are shown. For a Vs- range of 7.0 to
7.2 with a temperature of 131°F, concentration of acetic acid
in the vapor phase can be reduced as low as 3 parts per
million. Encouraged by the above test results, we are
planning to prove the operability and economy of this lime
process by 3,000 st.cu.ft./min. pilot plant in the near future
3. DESULFURIZATION OF FLUE GAS FROM COAL FIRED BOILER
Technogical requirements for the desulfurization of
flue gases from coal fired boilers are increasing from the
standpoint of environmental protection especially in other
countries than in Japan such as in the United States or in
Germany, where they have big coal reserves as an. energy
source for power generation. Also, an increase in power
generation from imported coals can be also anticipated in
the future in Japan.
In respect to these considerations, we have made a train
of desulfurization tests for flue gas from coal fired boiler
at the 60 st.cu.ft./min. test facility using sodium acetate-
lime stone process in order to solve problems, which may
encounter in the desulfurization of flue gases from coal
fired boilers. These tests were carried out in August of
1975, and the results are reported in the next section.
3-1. Sodium Acetate-Lime Stone Process
Following test items have been considered for the
evaluation of the process.
524
-------
1. Efficiency of desulfurization.
2. Efficiency of particulate removal.
3. Quality of gypsum.
4. Construction material.
5. Material and heat balance.
6. Behavior of trace materials and their accumulations in
the system.
7. Confirmation of the system for a closed loop without
disposal water.
A flue gas containing 600 parts per million of S02 and
1.4xlO~5 to 2.8xlO""5 gr./st.cu.ft. of particulate was
introduced into scrubber with an inlet gas temperature of
above 100°C.
As to the efficiency of S02 removal, nearly complete
removal of S02 has been achieved as listed in Table 3.
Accordingly, as far as efficiency of desulfurization
is concerned, the experimental results are excellent.
As to the removal of particulate in the flue gas, tests
have been made for variable contents of particulate and
different numbers of perforated plates in the scrubber.
The results are shown in Table 4.,
525
-------
In these tests, flue gas containing a. high level of
particulate (more than 0.14 gr./st.cu.ft.) has been prepared
by adding fly ash to the flue gas to determine the removal
efficiency, (s. Exp. No. 9 and 10 in Table 4.)
As seen in Table 4, efficiency of particulate removal
is remarkable even to the high contents of particulate and
a distinct relationship can be seen between particulate
removal efficiency and the number of perforated plates used.
As to particulate elimination, better results can be
anticipated for the scrubber with 12 perforated plates than
with an electrostatic precipitator.
During the above tests, no noticeable troubles have
been experienced with particulate or chlorides even in the
case of high particulate loadings.
The quality of gypsum formed was satisfactory for
further processing, even if it had colored gray to light
brown depending on the operational conditions.
No influence has been observed in the crystal growth
of gypsum, size of which ranged from 100 to 150 microns in
the long axis and 50 to 60 microns in the short axis.
As to materials of construction for the test, only
corrosion-resistant metals and plastic materials were used,
so that no abnormality has been observed during the tests.
Regarding the material and utility balance of this
process, a comparison was made between designed values and
experimentally obtained values using the 60 st.cu.ft./min.
526
-------
bench scale test facility. Comparative data are listed in
Table 5.
As can be seen from Table 5, smaller experimental
values than designed have been obtained as the material
balance for lime stone and gypsum, which can probably be
ascribed to smaller S02 content in the flue gas than
designed at the early stage of the test. Otherwise, an
excellent coincidence between experimental and projected
values have been attained both for material and utility
balance.
Regarding impurities such as particulate and chlorine,
tests have been made extensively. Acid insolubles in the
gypsum were analyzed and found to be composed mainly of
silica and alumina, which amount to about 4$ of the gypsum.
This suggests that the particulate in the flue gas is
removed from the system together with gypsum. Thus, in
our tests no build-up of particulate has been observed
during operation.
As to chlorine content in the flue gas, measurements
showed chlorine content of 0.7xlO~5 to l.0x!0"5gr./st.cu.ft.
No operational troubles have been caused by this level of
chlorine content, and also, no influence has been observed
upon the absorption of 302 and acetic acid in the scrubber,
the rate of oxidation and of gypsum formation. However, if
a closed system is adopted, accumulation of chlorine will
take place in the circulating liquor, and must be eliminated
from the circulating system. Extensive research is now
being made by us to solve this problem.
527
-------
As far as this bench scale test was concerned, a
complete closed loop has been successfully demonstrated with-
out waste water disposal. In this bench scale test, make up
of acetic acid and sodium hydroxide were not necessary.
Besides the above items mentioned, acetic acid vapor
in the flue gas from the S02 scrubbing section has been
measured both by gas chromatography and by gas detection
tube. The former method gave an acetic acid concentration
of 1.9 to 3-5 parts per million in the outlet gas of the
scrubber, while the concentration was Less than i part
per million for the same samples by the latter method,
showing practically complete removal of acetic acid vapor
in the flue gas.
As to pressure drop, a total pressure drop of 11 in.
H^O has been recorded for this process, which corresponds
to a mean drop of 2.17 in. f^O for each chamber of the
scrubber. (The scrubber consists of 2 chambers of SOp
absorption and 3 chambers of acetic acid elimination).
For each perforated plate, a mean pressure droi> amounts
to 0.51 in. H20. This magnitude of pressure drop for
perforated plate can guarantee good distribution of flue gas
in the large scale scrubber. This was confirmed by another
separate experiment by using test facility of 18,000 st. cu.ft,/min.
gas flow rate. Smooth flow down of the liquor and good
dispersion of flue gas into the counter-current liquor through
the perforated plates were observed in the above test.
To determine the possible trouble in the acetic acid
eliminator by the mist of the absorbent from the absorption
528
-------
section, careful observation has been made, but no
troubles have been found.
Scaling in the scrubber was prevented effectively by
the measures previously noted, and therefore no scaling
problem has been raised during the test.
Through a series of tests made on the flue gas from
coal fired boiler, we have succeeded in collecting necessary
information for the scale-up of this process, and it will be
soon demonstrated to verify the process by 3,000 st.cu.ft./min.
pilot plant, followed by demonstration of full scale plants.
4. ECONOMY OF THE PROCESS
On the basis of our experience and on a series of
bench and pilot tests, process economics can be estimated
with good accuracy.
In Table 6, a standard desulfurization cost of sodium
acetate-lime process for the flue gas rate of 588,500
st.cu.ft./min. is listed for 1976 costs in Japan. This
cost should be considered as a measure of our sodium
acetate - lime process, because it is strongly dependent
on the impurity content of raw materials and on the dealing
of gypsum, whether it is disposed or utilized. In case
gypsum is thrown away, process can be simplified and both
the investment cost and utilities cost can be reduced,
resulting in the reduction of desulfurization cost.
It should be further noted, that desulfurization
cost is greatly influenced by local conditions. However,
529
-------
as can be deduced from Table 6, this sodium acetate-lime
process is low in desulfurization cost and excellent in
operability.
5. SIMULTANEOUS REMOVAL OP S02 AND NOX BY SODIUM
ACETATE-GYPSUM PROCESS.
Taking advantage of sodium acetate-gypsum process, we
have further developed a process to eliminate S0« and NO
^ .A.
simultaneously. NO removal is furnished in the above
system by the addition of a catalyst, which is commonly
available in the market for a reasonable price.
In this process, NO is.reduced in the presence of the
catalyst to sodium imidodisulfonate by the sodium sulfite
in the absorbption liquor, and sodium imidodisulfonate
formed is further decomposed by calcium nitrite and lime in
the presence of sulfuric acid to nitrogen and gypsum.
As shown in Figure 5» a part of mother liquor containing
sodium imidodisulfonate, is drawn as a side stream from the
recirculation line to the scrubber and then, it is decomposed
by the above chemicals to nitrogen and gypsum in the reaction
vessel.
Accordingly, this system can eliminate S0~ and NO
£, -A.
simultaneously by additional attachment of NO removal
section to the sodium acetate-gypsum process in a complete
closed loop without no disposal water.
This process has been proven by 3000 st.cu.ft./min.
pilot test and based on this successful results, we are
planning to build a-full scale plant in the near future.
530
-------
6. CONCLUSION
Based on long experience of research and development
in the field of flue gas desulfurization, we have finally
come to sodium acetate-lime process for the treatment
of flue gases from both coal and oil fired boilers, process
economics can be assumed to be the most advantageous one.
Furthermore, we have recently developed a simultaneous
removal of SOp and NO process on the basis of the sodium
£— A.
acetate-gypsum process.
Though these new systems are still on the way of pilot test,
we are sure that these systems will be demonstrated in a full
scale plant in the near future as one of the most favorable
process for various type of flue gases containing S0? and
particulate.
531
-------
References
1) T. Xauiinaga and K. Noguchi
" Kureha Sodium Acetate-Gypsum FGD Process and Plant "
Chemical Factory (Japan) 19, 5, 85-88 (1975)
2) S. Saito
" Kureha Sodium Acetate-Gypsum FGD j rocess "
Chemical Engineering (Japan) 20, 3, 17-19 (1975)
532
-------
Table 1, SOp REMOVAL EFFICIENCY BY SODIUM ACETATE-GYPSUM PROCESS
AT 3000 st.cu.ft./min. PILOT TEST FACILITY.
C/l
C/4
Gas Flow Hate
Exp.No. (st.cu.ft./min.)
1
2
3
4
2940
2350
1470
940
S02
Concentration
Inlet (parts/mill ion)
1500 -
1500 -
1500 -
1500 -
1700
1700
1700
1700
Outletl
less
less
less
less
SOp Removal
'parts/million) Efficiency
than
than
than
than
1
1
1
1
more than
more than
more than
more than
99.
99.
99.
99.
9
9
9
9
-------
Table 2, LIME STONE AND GYPSUM QUALITY
Lime Stone
Particle Size
200 mesh pass
Chemical Composition
Si0
Gypsum
CaO 54.
MgO (as Mg) 0.3
Free Water 5.59^ (Wet base)
Crystal Water 20.755?- (Dry base)
CaO 30
SO, 45.
Purity 96.88^- (as CaS04.2H?0)
Crystal Size 50 ;i x 100 ji
534
-------
Table 3, REMOVAL EFFICIENCY OF S02 BY SODIUM ACETATE-LIMESTONE PROCESS
IN 60 st.cu.ft./min. BENCH SCALE TEST FOR FLUE GAS FROM
COAL FIRED BOILER.
Exp. No.
1
2
3
4
5
6
Concentration of c
Gas Inlet
646
690
570
550
570
580
jOp (Darts/million)
Gas Outlet
0
1
0
1
0
2
Desulfurization
100.0
99.8
100.0
99.8
100.0
99.6
-------
Table 4, RELATIONSHIP BETWEEN NUMBERS OF PERFORATED
PLATES AND PARTICULATE REMOVAL EFFICIENCY FOR A
FLUE GAS WITH A FLOW RATE OF 60 st.cu.ft./min.
Exp.No.
1
2
3
4
5
6
7
8
9
10
11
No. of
Plates
Used
3
3
3
5
5
12
12
12
12
12
12
Amount of Particulate
( gr./st.cu.ft. x 103)
Inlet
1.50
2.66
2.95
3.31
3.28
3.93
5.83
4.50
2.71
209.1
205.6
Outlet
0.90
1.02
1.16
0.88
0.85
0.48
0.51
0.23
0.28
2.04
1.47
Efficiency
of Removal
(*)
40
62
61
74
74
88
91
95
90
99
99
536
-------
Table 5, I'lATERIAL AND UTILITY BALANCE OF SODIUM ACETATE-LIMESTONE PROCESS
Gas Composition
S0? Inlet
S0? Outlet
Acetic Acid Outlet
Haw Material
0-J
By product
Gypsum
Utilities
Process tfater
Steam
Power
Air
Experimental
580-630 parts/million
0-2 "
1.6-3.5
0.55 lb/hr.
0.88 lb/hr.
0.42 gal./hr.
44.1 lb/hr.
10 kWh
141.2 st.cu.ft/hr.
Designed
600 parts/million
1 "
3
0.59 lb/hr.
1.01 lb/hr.
0.42 gal./hr.
176.3(max.)-22.0(min.)lb/hr,
40 kWh
141.2 st.cu.ft./hr.
-------
Table 6, ECONOMIC EVALUATION OP SODIUM ACETATE-LIME
PROCESS
Flue Gas Volume: 588,500 st.cu.ft./min.
S02 Content of Flue Gas: 1,000 parts/million
Efficiency of Desulfurization: more than 99^
Operating Hours: 8,400 hrs./year.
Investment Cost ,(BL): 15,800 ¥/KW.
Cost of Desulfurization: 12.57 ¥/gal.
Raw Materials and Utilities
Hydrated Lime: 7,429.5 Ib/hr.
NaOH : 35-3 Ib/hr
Acetic Acid: 201.6 Ib/hr.
Steam: 8,766.2 Ib/hr.
Process Water: 13,210.0 gal/hr.
Electricity: 4,500 KW
Gypsum: 172,840 Ib/hr.
Man Power
2 Men/Shift, 4 Shifts/day + 2 Men.
538
-------
C/n
W
1X5
QH
Keheater
Acetic Acid
Recovery
>>ect ion
:at i on
'-T-t:on
Flue Gas
3_J
Acetic Acid
Recovery T-mk
Absorbent Tank
Absorption
Tower
Oxida-
t ion
Tower
—Air
Gypsum
Recovery
Reactor
Separator
•Lime Stone
•Water
Slurry Tank
Mother Liquor Tank
FIGURE 1, SCHEMATIC FLOW DIAGRAM OF LIME STONE DESULFURIZATION PROCESS
-------
FIGURE 2, SOLUBILITY OF CALCIUM SULFATE IN THE ABSORPTION
LIQUOR IN DEPENDENCE ON SODIUM SULFITE CONCENTRATIONS
o
3
&
•H
O
CO
C6
O
a
c:
o>
o
c
o
o
0.5
0.4
0.3
0.2
0.1
0
Chemical Composition of Liquor
CH5COONa
Na2S04
CHjCOOH
CaS04•2H?0
0
1
Concentration of
540
-------
Acetic Acid
Recovery
Section
Desulfuri-
zation
Section
Fan
FIGURE 3, LIME DESULFURIZATION PROCESS
Ca(OH).
Oxidation
Tower
Liquid
Cyclone
fcQ
Separator
A N
—Q
Gypsum
Recovery
Reactor
ta.
Absorption
Tower
Air
Compressor
Gypsum
Mother
Liquor
Tank
-------
FIGURE 4, RELATIONSHIP BETWEEN P11 OP THE LIQUOR AND
ACETIC ACID CONCENTRATION IN THE VAPOR PHASE,
o
•H
8
m
-p
0.
O
O
O
O
C!
O
-H
-P
£0
f-t
-p
A
0>
o
fl
o
o
Q)
CO
OJ
^3
04
0)
ce
Ctl
CH^CQONa SOLUTION
1 -
of Solution
542
-------
FIGURE 5, SIMULTANEOUS REMOVAL OF S02 AND NOX
BY SODIUM ACETATE-GYPSUM PROCESS.
On
A
CaCO3
Acetic Acid
Recovery
Section
SOX and NOX
Absorption
Section
H20
Flue Gas
Air
Oxidation
Tower
1 ------ -
Ca(OH)2
Ca(N02)2
H2S04
CaCO,
Steam
Nitrogen
Fo rmi ng
Gypsum Recovery
Reactor
Absorption
Tower
-------
THE BUELL DOUBLE-ALKALI SO CONTROL PROCESS
H. Edward Bloss
Buell-Envirotech
Lebanon, Pennsylvania
James Wilhelm
EIMCO-Envirotech
Salt Lake City, Utah
William J. Holhut
Central Illinois Public Service Company
Springfield, Illinois
ABSTRACT
This paper traces the development of the Buell Double-Alkali SO,
Control Process to the current technology of concentrated and dilute
mode operation that are currently offered commercially. Areas of
application are defined for the two modes, each resulting in elimination
of chemical scaling, abrasive scrubbing slurries and high maintenance
expenditure. Design information is also presented for utilities
service with high chloride coals. The Central Illinois Public Service
Company, Newton Station, Unit #1 (concentrated mode, high chloride
pulverized coal operation) currently under design is presented and
equipment defined. Preliminary capital investment and operating
cost estimates are tabulated.
545
-------
INTRODUCTION
Scrubbing of sulfur dioxide from boiler flue gas has been
a major problem for the utilities industry due to costs,
complexities, and impact on reliability of electric gener-
ating operations. An important solution to this problem
lies in continued advancement of process design technology
which will achieve required reduction in stack S02 emission
without reducing generating unit availability and without
creating secondary environmental pollution effects that
may jeopardize the quality of surface and ground water
bodies.
Because of variations in boiler installations and fuel
sources, a single scrubbing system design has not been
able to achieve the objectives of all applications. For
that reason, Envirotech has invested more than four years
work in developing and evaluating the applicability of
new process design technology to control the various
operating variables encountered with operating boiler
units.
ENVIROTECH'S APPROACH
As a means of minimizing capital investment, the utilities
industry continues to focus on flue gas desulfurization
systems of the waste "throwaway" type. In view of their
comparatively low cost and their ability to precipitate
collected sulfur oxides as a water-insoluble solid,
calcium-containing raw materials such as lime or limestone
are solely used to convert the collected S02 waste to the
solid form. The basic double-alkali system shown on
Figure 1 accomplishes the above in two advantageous steps,
scrubbing first, post-precipitation later:
Equation 1 (Absorption)
S02 + Na2SO3 + H2O 2 NaHSO3
Equation (la.) SO2 + H2O H2SO3
Equation (Ib.) H2S03 + Na2S03 2 NaHSO3
and/or
Equation 2 (Absorption)
S02 + H20 + 2 NaOH Na2S03 + 2H20
546
-------
Equation 3 (Regeneration)
2 NaHSOs + Ca(OH)2 Na2SO3 + CaSOo.1/2 H90
+3/2 H20
and/or
Equation 4 (Regeneration of fully oxidized collected
sulfur oxides)
Na2SO4 + Ca(OH)2 + 2 H2O CaS04.2 H2O + 2 NaOH
All sodium-base double-alkali systems perform generally
within these equations.
Beginning with initial work five years ago, Buell and Eimco
have committed efforts to the advancement of the double-
alkali design approach to SO2 control technology. The
throwaway SO2 system double-alkali has a special advantage
in avoidi-ng precipitation of the SO2 solids in the gas-
handling scrubbing system and instead, applies the lime
treatment externally of a clear liquor scrubbing circuit
that utilizes highly water-soluble sodium chemicals as
shown in equations above.
Envirotech's work in double-alkali flue gas desulfurization
has developed two similar flow sheets that are now being
offered or applied in commercial applications. The first,
the dilute mode process - is specifically designed for
boilers operating with abnormally high excess air or for
low sulfur coal applications such as in western U.S.A.
The second, the concentrated mode process - is used for
high sulfur coal applications in pulverized coal fired
boilers where the excess air in the flue gas is generally
low enough to limit the oxidation of the absorbed sulfur
species, as required for its zero-effluent operation.
DILUTE MODE PROCESS
For power plants that have high excess air and/or low
flue-gas S02 concentrations, the oxidation rate of the
absorbed SO2 changes the chemical equilibrium relation-
ships in the system. The major alkali in the process is
sodium hydroxide not sodium sulfite and the calcium ion
concentration in the regenerated liquor is as high as
500 PPm. in order to control chemical scaling of the
scrubber due to intrusion of calcium via the regenerated
liquor return stream, a calcium softening step is utilized
as shown in Figure 2. Soda ash makeup required by the
system is added to this softener, or Reactor-Clanfier,
547
-------
to precipitate soluble calcium and eliminate scaling
potential in the scrubber. The liquor flow volumes in
the post-precipitation step of this system are dictated
by the concentration of alkali (NaOH) that can be formed
in the regenerated liquor under conditions of chemical
equilibrium governed by Equation 4, above.
The normal ionic concentrations for this system are 0.1
molar NaOH and 1 molar Na. This variation of the basic
double-alkali system provides scale control and process
reliability but has a captial cost higher than a compar-
able concentrated mode (if it were applicable), due to
the equipment for control of the soluble calcium and the
higher volume of liquor that must be handled by the
post-precipitation system.
CONCENTRATED MODE PROCESS
In the concentrated mode (concentrated active-alkali)
process, the flue gas is scrubbed with a concentrated
sodium sulfite/bisulfite solution to remove sulfur
oxides. A bleed stream of spent scrubbing liquor at a
pH near 6.0 is reacted with lime to precipitate calcium
sulfite and the limited amount of calcium sulfate that
has been formed due to oxidation in the scrubber. The
precipitated solids are then thickened, filtered and
water-washed (for minimization of sodium losses), to
produce a disposable waste for landfill disposal, and
the regenerated liquor is returned to the scrubber to
absorb additional SO2. The normal ionic concentrations
for this system are approximately 0.5 molar Na2S03 and
2 molar Na. This flow sheet is shown in Figure 3.
Due to limitations in the calcium sulfate precipitation
kinetics, this system is used where the oxidation of
the S02 may be kept below about 25% of the total absorbed
SOX species. The magnitude of this oxidation rate depends
upon the ratio of the S02 absorbed to the excess air or
oxygen in the flue gas. In general, this ratio must be
at least 20 ppm of SO2 per excess air percentage in order
to adequately control the oxidation rate. As a typical
example, a boiler operating at 30% excess air requires a
minimum of 600 ppm or about 1% sulfur in the coal to
insure that oxidation levels are sufficiently limited. At
sulfur levels greater than the 600 ppm in gas, or 1% in
fuel, effective precipitation of the sulfate ions is assured.
548
-------
This double-alkali system removes at least 90 to 95%
to the SC>2 in the flue gas. A relatively dry filter
cake waste is the sole waste product. There are no
liquid effluents discharged from the process. Scrubber
scaling is controlled because the regenerated liquor
contains a very low calcium ion concentration (20-80 ppm)
because of the limited aqueous solubility-product con-
stant of the calcium sulfite species. The scrubbing
solution is a clear liquor (containing only a small
amount of residual fly-ash), that has essentially no
abrasive characteristics and cannot produce internal
scaling.
HIGH CHLORIDE PROCESS
Absorption of hydrogen chloride from the flue gas can
cause special problems in all throwaway systems including
the double-alkali process. As indicated above, filter
cake washing is employed to wash out and recycle the
soluble sodium salts to the process. This procedure
minimizes the amount of soda ash required for make-up.
When chloride'is absorbed into the scrubbing liquor,
sodium chloride concentration-buildup in the system
is magnified by this washing and recovery of the soluble
salts from the filter cake. In some cases, augmental
sodium chloride concentrations as high as 1.5 to 2.0'
molar may result, causing sodium sulfate insolubility
problems as well as increased usage of make-up soda ash,
with higher sodium content in the filter cake.
In order to eliminate these problems, a modification to
the basic flow sheet has been developed to remove the
chlorides in a separate precooler section upstream from
the S02 absorber. This flow sheet is shown in Figure 4.
The chlorides and other secondary coal contaminants are
removed in a low pH scrubbing liquor. A slipstream of
the resulting acidic solution is neutralized with lime
or limestone to produce a calcium chloride solution which
is used as a final wash on a horizontal fliter of the
Eimco-Extractor type. This wash is so designed that the
filter cake absorbs the chlorides.
The net results of this chloride removal loop when applied
to h?gh-chloride coals (approximately 0.1% chloride by
weight in coal or higher), are to:
1. Reduce sodium carbonate make-up requirements;
thus reduce operating costs to normal double-
alkali system levels.
549
-------
2. Reduce the dissolved solids concentration
in the SC^ scrubbing system; thus to reduce
salting out problems and achieve normal
optimum double-alkali process operation.
3. Reduce the amount of sodium salts in the
filter cake; thus producing more suitable
landfill waste material.
4. Eliminate most of the residual fly-ash
and other secondary coal contaminants from
the SO2 removal system. This minimizes
the amounts of heavy metal ions in the S02
scrubbing liquor and thus reduces the
catalytic effects of these ions in promoting
oxidation of the sulfite and bisulfite ions.
In a further refinement of this process, a fly ash thick-
ener is added to the precooler loop. Thus, the system
can be used as a back-up to the electrostatic precipitators
and still operate at a low suspended solids level with low
abrasion characteristics in the precooler liquor. The
primary function of this additional thickener is to accommo-
date temporary malfunctions of the electrostatic precipitators
such as shorts of electrical sets and temporary power loss.
In the interim, while the back-up function is actually being
fulfilled, the precooler system is not subjected to abrasion
since solids build-up and related problems have been avoided:
the system will continue to function as designed. This is
the basic flow sheet that has been accepted and is being
applied for use at the Newton Station of Central Illinois
Public Service Company, scheduled for 1977 startup of
operations. The system, currently in the engineering
phase, is the concentrated mode/high chloride process
with auxiliary fly-ash thickener loop to control the system
during precipitator upsets.
PLANT DESCRIPTION & DESIGN CONDITIONS
The double-alkali system will be installed as an adjunct
to the Newton Station, Unit #1, Newton, Illinois. The
boiler is a drum-type manufactured by Combustion Engineering,
Inc., designed for a maximum steam generating capacity of
4,158,619 Ibs/hr. at 2,620 psig and 1,005°F. at the steam
outlet connection. The boiler unit and the double-alkali
control system are scheduled for in-service operation
December, 1977.
550
-------
The boiler is tangentially fired with pulverized coal.
The pulverizing mills are the bowl-type manufactured
by Combustion Engineering, Inc. Net theoretical air
required for combustion at maximum steaming condition
is 4,192,000 Ibs/hr at 60°F. and 80% RH. The boiler
is designed for operation at 22% excess air. High
sulfur Illinois bituminous coals from several sources
may be burned. In the design of the system, the
following flue gas analysis was developed from the
worst case criteria:
LBS/HR.
S02 37,800
H2O 330,000
02 556,000
N2 4,580,000
C02 1,110,000
S03 700
HC1 1,000
TOTAL 6,615,000 d)
The system is designed for coal properties and boiler
system characteristics as follows:
Sulfur Content 4.0% in coal
Chloride Content 0.2% in coal
Coal Heating Value as received 10,900 BTUS/lb.
Boiler Heat Release 5.5 x 109 BTUS/hr,
95% of Sulfur in fuel Reports as S02 in flue gas
96% of Cl in fuel Reports as HC1 in the flue gas
represents total I.D. fan capacity including excess
and in-bleed air leakage.
551
-------
PROJECT SCOPE
Buell is responsible for the design, procurement and
delivery to the job site, of a double-alkali flue gas
desulfurization facility (FGDU) designed for 90%
removal of S02- The station design includes two
induced fans which discharge into a common plenum.
Each fan is also connected to the stack. The double-
alkali flue gas desulfurization unit, which is being
built to accommodate Unit 1, will include the following
major equipment and facilities:
A. Four booster fans.
B. The ductwork, including the plenum, leading
to the four booster fans, also including
isolation dampers for each fan.
C. Isolation gate dampers in the boiler ductwork
to isolate the I.D. Fans from the stack.
D. Four precooler towers.
E. Ductwork from the booster fans to the four
precoolers.
F. Four scrubber towers.
G. Mist eliminators between the precoolers
and the scrubbers.
H. Mist eliminators at the outlet of the four
scrubber towers.
I. Ductwork from each of the four scrubbers con-
necting them to two plenums, each of which
is common to two scrubbers. This ductwork
will be complete with outlet isolation gates
for each scrubber.
J. All ductwork to connect the scrubber system
back to the chimney.
K. Lime unloading and handling system including:
1. Equipment designed to receive pebble
or granular"quick-lime using both
rail and truck unloading facilities.
552
-------
2. Four lime storage silos for the
storage of pebble or granular lime.
3. Lime feeding and ball-mill-slaking
facilities for the addition of lime
to the reaction tanks.
4. A building sheltering the lime feed-
ing and slaking facilities.
L. Soda ash system, consisting of:
1. Soda ash off-loading facilities de-
signed to receive soda ash from both
rail and truck.
2. Two soda ash storage silos.
3. Soda ash feeding and dissolving equip-
ment.
M. Two 100' diameter thickeners for the double-alkali
post-precipitation system with concrete bottoms and
access to bottom discharge cone using tunnel.
N. One 50" diameter thickener for the precooler liquor
loop, constructed of coated steel and elevated above
grade.
O. Three Eimco Extractor filters for dewatering and
washing of double-alkali-system thickener underflow,
and associated equipment including vacuum pumps,
filtrate receivers, filtrate pumps, and moisture
traps. A building to house the filters and all
related equipment is also included with the FGDU.
P. A control and maintenance building.
Q. Flocculant addition systems for the thickeners.
R. Material off-loading facilities including a train
shed (about 250' long) which will accommodate
four cars and material handling equipment to
unload two cars at one time from a total of six
hoppers. The material off-loading facility is
able to handle trucks which may or may not have
their unloading auxiliaries.
The Project Scope also includes necessary accessary equip-
ment and facilities required to fulfill all performance
guarantees.
553
-------
GENERAL COMMENTS
Design of flange-to-flange wet-end equipment such as
reactors, thickeners, vacuum filters, is based on the
process bleed liquor flow rates in accordance with
established sizing criteria per Envirotech pilot plant
experience. Wherever feasible, the Envirotech wet-end
system design incorporates gravity flow for operating
simplicity and to reduce oxidation, covered troughs
have been utilized in place of pipe for gravity runs
wherever possible. Materials of construction are
generally mild steel lined with appropriate corrosion
and/or abrasion resistant materials. Piping is sized
for flow rates of 4-7 feet per second. Process pumps
are designed per system flow sheet with an added 20%
safety factor. Rubber lined pumps with rubber-covered
impellers are utilized for corrosive and/or abrasive
service. All process pumps are spared where necessary
to insure continuity of system operation.
Instrumentation is designed to permit automatic oper-
ation of the system, and includes compensation for
load changes, and- an emergency water deluge system for
high-temperature protection of scrubber internals
including non-metallic linings. Facilities provide
for the recording of key system operating data, i.e.,
liquid flow rates, temperatures, pressures, pH, inlet
and outlet SC>2 concentrations, etc. Manual override
is also provided for. In the event of complete system
shutdown, provisions have been made to drain all lines
and pumps to drain sumps. The drain sumps are then
pumped to vessel freeboard space in the post-precipita-
tion system. Tankage is sized to allow for the minimum
liquid retention time consistent with optimum process
operation taking into account the additional volume
needed for liquid storage during system shutdown.
Materials of construction are generally non-metallic
lined, 1/4" thick, mild steel. A vacuum filtration
system is provided and is sized based on the thickener
bottoms solids-flow rate. Provision for a spare filter
plus variable speed drive permits extra capacity.
Should it become necessary to remove the filters from
operation for a short period of time, the thickeners
are capable of handling an amount of solids build-up.
The thickener mechanisms are designed to permit an
accumulation (12" - 24") of settled solids in the
thickener tanks through the incorporation of automatically
operated lifting devices that raise and lower the rakes.
554
-------
The wet-end (post-precipitation) system incorporates
a filter building which houses the vacuum filter system.
A separate building houses the process control room.
The latter is a dust free, heated and air conditioned
room housing the system instrumentation and operators'
comfort facilities. The reaction tanks are provided
with a launder system that permits by-passing of any
one unit for maintenance without shutting down the
process.
CAPITAL INVESTMENT AND ANNUAL OPERATING COST ESTIMATES
The project is currently in the design engineering phase
and finalized costs have yet to be developed. It is
currently estimated that the direct investment will be
$23 million with the completed cost of $40 million which
would reflect the total capital investment estimated at
this phase of the project. Since the investment costs
have not been developed at this time for all the indiv-
idual equipment items, an analysis of the fixed invest-
ment cannot be made. However, based on the estimates
and the known feedstock/utility and manpower requirements,
the following annual operating cost values have been
developed:
ESTIMATED AVERAGE ANNUAL OPERATING COSTS
DOUBLE-ALKALI PROCESS
(575 MW NEW COAL FIRED BOILER UNIT, 4.0% IN FUEL; 90% REMOVAL)
DIRECT COSTS
ANNUAL
QUANTITIES
UNIT
COST $
TOTAL
ANNUAL
COST $
Lime (93% Avail.
Ca0) 114,975 M 35/ton 4,024,125
tons
3.833 M
tons
Subtotal Raw
Materials
80/ton 306,640
4,330,765
PERCENT OF
TOTAL ANNUAL
OPER. COST
30.9
2.3
33.2
555
-------
TOTAL PERCENT OF
ANNUAL UNIT ANNUAL TOTAL ANNUAL
DIRECT COSTS QUANTITIES COST $ COST $ OPER. COST
Conversion Costs
Operating Labor
and Supervision 42,924 Man 8.00 343,392 2.6
Hours per
Man Hrs.
Utilities
Steam 53,300 M 1.20/M 183,960 1.4
Lbs. Lbs.
Process Water 312,732 M 0.30/M 93,820 0.7
Gal. Gal.
Electricity 79,716,000 0.012 956,592 7.3
KWH per KWH
Maintenance
Labor and Material
(.02 x 40,000,000) 800,000 6.1
Subtotal Conversion
Costs 2,377,764 18.3
Subtotal Direct Costs 6,708,529 51.5
556
-------
PERCENT OF TOTAL
INDIRECT COSTS TOTAL ANNUAL COST AMOUNT OPER. COSTS
Average Capital Charges
@ 14.0% of Total
Capital Investment 5,600,000 43.0
Overhead
Plant
683,553 5.3
Administrative, 10% of
Operating Labor 34,339 0.2
Subtotal Indirect Costs 6,317,892 48.5
Total Annual Operating
Cost 13,026,421 100.0
DOLLARS/TON CENTS/MILLION DOLLARS/TON
COAL BURNED MILLS/KWH BTU HEAT INPUT SULFUR REMOVED
Equivalent Unit
Operating Cost 8.42 3.69 38.62 233.89
ESTIMATE BASIS;
Preliminary Project Cost Estimate
Life of Boiler Plant: 30 years
Coal Burned, 1,547,064 tons/year, 10,900 BTU/hr.
Stack Gas Temperature -154°F. (Combination Steam Reheat
& Bypass Mixing)
Boiler Unit on Stream Time, 6,132 hrs. @ Full Load -
(70% Load Factor)
Total Capital Investment, $40,000,000 (est'd.); Subtotal
direct investment $23,000,000 (est'd.)
557
-------
CONCLUSION
In order to achieve a flue gas desulfurization system
design that results in a high degree of reliability for
a specific boiler application, recognition of all key
criteria and variables is required. The double-alkali
system operating in the concentrated mode is a system
design with no scaling potential and minimized abrasion
when applied to a modern pulverized-coal utility boiler
service. With low-quality coals, with high chloride
content, the addition of the separate precooler absorp-
tion loop for chloride as well as fly-ash control is a
further development for that increased reliability and
reduced operating cost.
These systems currently have redundancy and/or excess
capacity in their design to accommodate operating
unknowns. As technology develops, the redundancy can
be reduced and the complexity capacity reduced. With
this development, investments cost and operating costs
will be reduced and reliability will continue to be
achieved.
558
-------
CLEAN GAS
SO 2 LADEN
FLUE
}
GA3^
makeup wate
clear liquor
^overflow
'
S02
ABSORPTION
r
ti
,
h
REGENERATION
a
SEPARATION
1
^
^^
_* regenerated sodium liquor
i
lime
1 dewatered waste solids ^
to disposal
If
REGENERATED
LIQUOR
INVENTORY
t
_ soda ash
, **
fig. I - BASIC SCHEME
-------
HOT FLUE 6AS
PRECOOLER SCRUBSER
SCRUBBED GAS
lime
STACK
u compressed _
-^- -o soda ash
REACTORcLARIFIER
VACUUM! FILTER
fig. 2 - DILUTE MODE
-------
Ul
HOT FLUE GAS
i—O—
SCRUBBED GAS
PRECOOLER SCRUBS
ER
lime
STACK
REACTION TANKS
soda ash
fig. 3- CONCENTRATED MODE
-------
HOT FLUE GAS
Ln
ON
K)
SCRUBBED GAS
REACTION TANKS
soda ash
'*!'
THICKENER
NEUTRALIZED 'v
fig. 4- HIGH CHLORIDE MODE
-------
TECHNICAL REPORT DATA
(Phase read Jaitnicrions on the reverse before completing)
REPORT NO.
EPA-600/2-76-136a
TITLE AND SUBTITLE
'roceedings: Symposium on Flue Gas Desulfurization-
ew Orleans, March 1976; Volume I
. RECIPIENT'S ACCESSION-NO.
. REPORT DATE
May 1976
. PERFORMING ORGANIZATION CODE
AUTHOR(S) .
R,D. Stern, Chairman; W. H. Ponder and
.. C. Christman (TRW, Inc.), Vice-chairmen
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING OR9ANIZAT1ON NAME AND ADDRESS
Miscellaneous
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
In-house
2. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13 TYPE OF REPORT AND PERIOD COVERED
Proceedings; 3/8-11/76
14. SPONSORING AGENCY CODE
EPA-ORD
s. SUPPLEMENTARY NOTES jjgRL-RTP project officers for these proceedings are R.D. Stern
and W.H. Ponder, Mail Drop 61, Ext 2915.
6. ABSTRACT The proceedingS document the presentations made during the symposium,
which dealt with the status of flue gas desulfurization technology in the United States
and abroad. Subjects considered included: regenerable, non-regenerable, and
advanced processes; process costs; and by-product disposal, utilization, and
marketing The purpose of the symposium was to provide developers, vendors, users
and those concerned with regulatory guidelines with a current review o? progress
made in applying processes for the reduction of sulfur dioxide emissions at the full-
and semi-commercial scale.
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS^
Air Pollution
Flue Gases
Desulfurization
Sulfur Dioxide
Sulfur Oxides
Cost Effectiveness
Byproducts
Disposal
Utilization
Marketing
Air Pollution Control
Stationary Sources
COS AT I Field/Gioup
13B
2 IB
07A,07D
07B
14A
-------