EPA-600/2-77-242
December 1977
Environmental Protection Technology Series
FLY ASH CONDITIONING
WITH SULFUR TRIOXIDE
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental Protec-
tion Agency, have been grouped into nine series. These nine broad categories were
established to facilitate further development and application of environmental tech-
nology. Elimination of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interegency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumen-
tation, equipment, and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the new or Improved tech-
nology required for the control and treatment of pollution sources to meet environmental
quality standards.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved for
publication. Approval does not signify that the contents necessarily reflect the views and
policies of the Government, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/2-77-242
December 1977
FLY ASH CONDITIONING
WITH SULFUR TRIOXIDE
by
Edward B. Dismukes and John P Gooch
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
Contract No. 68-02-2114
Task No. 1
ROAP No. 21ADL-027
Program Element No. 1AB012
EPA Task Officer: Leslie E. Sparks
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
Table of Contents
Section Page
I Introduction and Summary 1
II Background 5
III Results from Test Program 7
IV Mass Balance Considerations 50
V Computer Model Projections of Precipitator
Performance 59
Acknowledgements 63
References 64
ill
-------
List of Tables
Table No. Page
1 Mass Train Results Reported by Research
Cottrell 16
2 Mass Train Results Reported by Kin
Associates 18
3 In situ Resistivity from George Neal
Plant, Unit 2 20
4 George Neal-Unit 2, Average voltages
and currents for 3/27/76 31
5 Average voltages and currents for
5/17/76 32
6 Ultimate Analyses of Coal Samples 34
7 Analyses of Flue Gas 35
8 Analyses of Hopper Ash 40
9 pH Values and Soluble S0i,~2 Concentra-
tions of Hopper Ash 42
10 pH Values and Soluble Sulfate Concentra-
tions of Cyclone Samples of Ash 46
11 pH Values and Soluble SO4~2 Concentra-
tions of Filter Samples of Ash 48
12 Injected S03 Found in Outlet Cyclone
Samples of Ash 57
13 Injected S03 Found in Outlet Filter
Samples of Ash 58
IV
-------
List of Figures
Figure No. Page
1 Inlet Size Distribution (Cumulative
Percent) for March 27, 1976 with 90%
Confidence Intervals ....................... 8
2 Inlet Size Distribution (Cumulative Mass)
for March 27, 1976 with 90% Confidence
Intervals .................................. 9
3 Average Inlet Size Distribution (Cumulative
Percent) for May 18-21, 1976 with 90%
Confidence Intervals ....................... 10
4 Average Inlet Size Distribution (Cumulative
Mass) for May 18-21, 1976 with 90%
Confidence Intervals ....................... 11
5 Outlet Size Distribution (Cumulative
Percent) for May 19, 1976 .................. 13
6 Outlet Size Distribution (Cumulative Mass)
for May 19, 1976 ........................... 14
7 Fractional Efficiency for May 19, 1976
(Based on single outlet sampling point).... 15
8 Resistivity Probe Voltage -Current
Characteristics without SO3 Injection ...... 21
9 Resistivity Probe Voltage-Current
Characteristics with S03 Injection ......... 23
10 Arrangement of Transformer Rectifier Sets
for George Neal Unit 2 ..................... 24
11 V-I Characteristics for TR Set 3 on
March 27, 1976 ............................. 25
12 V-I Characteristics for TR Set 4 on
March 27, 1976 ............................. 26
13 V-I Characteristics for TR Set 7 on
March 27, 1976 ............................. 27
14 V-I Characteristics for TR Set 8 on
March 27, 1976 ............................. 28
v
-------
List of Figures
(Continued)
Figure No. Page
15 V-I Characteristics for TR Set 3 on
May 21, 1976 29
16 V-I Characteristics for TR Set 8 on
May 21, 1976 30
17 Comparison of Predicted Gas Concentrations
(Functions of Excess Air) with Observed
Concentrations (Displayed by Dashed
Horizontal Lines) 36
18 Hopper Configuration for the Precipitator.. 38
19 Soluble SO.,"2 in Hopper Ash as a Function
of Hopper Location or Gas Temperature 43
20 Locations for Sampling with Series
Cyclones 45
21 Comparison of H2SCK Concentrations at the
Dew Point with Experimental Results 53
22 Relationship between Flue-Gas Temperature
and Experimental HaSOi, Concentrations 54
23 Computer Model Projections of Collection-
Efficiency without SO3 60
24 Computer Model Projections of Collection-
Efficiency with S03 62
VI
-------
I. INTRODUCTION AND SUMMARY
This report describes a study conducted by Southern Research
Institute for the purpose of evaluating an S03 injection system
for the George Neal Unit 2 Boiler of Iowa Public Service Company
in Sioux City, Iowa. The SO3 injection system was supplied by
Research Cottrell for the purpose of increasing the collection
efficiency of the electrostatic precipitator installed on the sub-
ject boiler. The study was sponsored jointly by Iowa Public Service
Company and the Industrial Environmental Research Laboratory of
the Environmental Protection Agency. Mass loading determinations
were conducted by Research Cottrell and Kin Associates. EPA's
portion of the study was concerned primarily with accounting for
the fate of the injected conditioning agent with emphasis on stack
losses of SO3, whereas IPS's objective was to determine whether
the injection system would provide a means of reliably increasing
the precipitator collection efficiency to 99% at full load with
normal plant operating conditions. A second objective of IPS was
to obtain an estimate of the specific collecting area which would
be required to achieve 99.0% collection efficiency without condi-
tioning when the precipitator is collecting fly ash produced from
the low sulfur, low sodium western coal which is used for fuel at
this installation.
The study consisted of a two-phase test program that was fol-
lowed by appropriate laboratory measurements and analyses. The
two phases of the test program were: (1) An evaluation of precipi-
tator performance at full load without S03 injection. These tests
were conducted on March 27, 1976. (2) An evaluation of precipita-
tor performance at full load with SO3 injection. These tests were
conducted during the week of May 17, 1976.
Results from the baseline tests on March 27 indicated that, as
expected, precipitator performance was limited by the high electrical
resistivity of the collected dust. The average collection effic-
iency reported by Research Cottrell was 91.3% at a plant load of
1
-------
299 megawatts. Measurements with a point-plane probe indicated a
dust resistivity value of approximately 6 x 10>2 ohm-cm at 118°C
(245°F). Voltage-current relationships obtained from the TR sets
also exhibited behavior typical of precipitators collecting high
resistivity dust. These voltage-current relationships were used
to estimate the allowable electrical operating conditions which
could be maintained without back corona or excessive sparking in
the absence of S03 conditioning. The estimated electrical operating
conditions, the precipitator geometry, and the measured particle
size distributions at the precipitator inlet were used as input
data to a computer program which simulates the operation of the
precipitator. The computer program was then used to estimate the
specific collecting area that would be required to achieve 99 0%
collecting efficiency. If it is assumed that ^ enlarged precipi_
tator should include a sufficient safety margin to allow about 12%
of the collecting area to be de-energized without decreasing per-
formance below 99% collection efficiency, the program output in-
dicates that a specific collecting area of 108 mV(mVsec) (550 ft2/
1000 ACFM) would suffice.
The test program with S03 injection was not conducted in ac-
cordance with our original test plan because of difficulties with
the SO, injection system and the precipitator TR sets "tripping
out it is our understanding that the cause of the difficulty
with the TR sets was ash build-up in the hoppers. As a result of
these problems, only one efficiency test at a load of 300 MW was
obtained with the S03 injection system operating continually and
without transformer-rectifier failures. The results obtained were
as follows:
Outlet Dust
Test at Stack
ASME 0.0508(.0222)
EPA 0.0703(.0307)
EPA with residue from
first impinger 0 .0828 ( .0362)
2
PreciPitator Efficient
9896
98>78
-------
These results were obtained from the Kin Associates report concern-
ing the subject test program, and the precipitator efficiencies are
based on an inlet dust concentration of 10.38 gm/DNm3 at 21°C (70°F)
(4.536 gr/DSCF). Although the above data indicate the performance
of the precipitator was approaching the desired value of 99%, addi-
tional test data taken after the injection system and precipitator
have operated under reasonably steady-state conditions for several
days would be required to determine if 99% collection efficiency can
be consistently maintained. This will require a solution to the ash
build-up problem and also a turn-down capability for the S03 system.
The reported rate of S03 injection during the test period was about
25 ppm by volume.
Our conclusions with regard to the effect of S03 injection at
this rate on the flue gas and fly ash properties, and the precipi-
tator performance, may be summarized as follows:
(1) Dust resistivity values measured with the point-plane probe
indicated that resistivity decreased to approximately 4 x 1010 ohm-
cm at 143°C (290°F).
(2) Voltage-current curves obtained from the precipitator power
supplies indicated that the dust resistivity was not limiting the
electrical operating conditions. The electrical operating conditions
with SO3 conditioning, the precipitator geometry, and the measured
inlet size distribution were used as input data to the mathematical
model. The results indicated that a collection efficiency of greater
than 99% is theoretically possible with a specific collecting area
of 39.37 m2/(nt3/sec) (200 ft2/1000 ACFM) and with the improved volt-
ages and currents.
(3) At the inlet of the precipitator, where the gas temperature
averaged 128°C (262°F), about 2 ppm of the added SO3 was found in the
gas phase, and about 23 ppm was found on the suspended fly ash. Ash
samples were collected and fractionated by size in a series of
cyclones; variations in the sulfate content of different samples
may have been caused by fluctuations in either boiler load (total
gas flow) or S03 injection rate. Ash of smaller particle size contained
-------
higher weight-percentage of sulfate than ash of larger size, as expected.
(4) In the precipitator hoppers adjacent to the center line
through the precipitator, about 15 ppm of the injected S03 was
found as sulfate on the ash. This quantity of S03 was lower than
expected from other data (S03 injection rate and S03 found as sulfate
on ash at the inlet and outlet). The discrepancy presumably was
caused by a lower-than-average rate of S03 injection in nozzles
directly upstream from the hoppers in question.
(5) At the outlet of the precipitator, where the gas tempera-
ture averaged 118°C (244°F), less than 1 ppm of H2SCK vapor was
found in the gas stream, and only about 2 to 3 ppm of S03 was found as
sulfate on the ash. The H2S04 vapor concentration was too low to
be distinguished from that occurring without conditioning. The
total sulfate on the outlet ash was only about 50% of the value
found without conditioning. Even though the weight fraction of
sulfate on the ash was increased by conditioning, the product of
(1) sulfate fraction .and (2) total ash concentration was lower with
conditioning than without, because of the marked increase in precip-
itation efficiency. In other words, the decrease in factor (2) as
the result of conditioning was more important than the increase in
factor (1).
In summary, the data indicate that an adequate accounting was
made for the fate of the injected S03 and that the overall rate of
SO3 emission from the stack (counting both E2SO^ vapor and sulfate
on the ash) was lower with conditioning than without. Additional
testing will be required to determine whether the injection system
can enable 99% collection efficiency of fly ash in the precipitator
to be reliably achieved.
-------
II. BACKGROUND
Collection efficiency measurements on the electrostatic pre-
cipitator installed on Unit 2 of the George Neal Station have in-
dicated that the precipitator performance is appreciably below the
design value of 99% at full load conditions. Although the precip-
itator performance has in the past reportedly been limited by factors
other than dust resistivity (specifically, a poor gas velocity dis-
tribution and hopper sweepage), both the electrical readings of
transformer-rectifier sets and in situ resistivity measurements have
shown that the electrical operating characteristics are severely
limited by dust resistivity.
In order to achieve the design value for collection efficiency
at this installation, two options are feasible: (a) increase the
plate area of the precipitator, and (b) lower the dust resistivity
to the extent that it does not limit the performance of the unit.
Option (a) involves a large capital expenditure since it is esti-
mated that the total plate area would have to be increased by a
factor of about 3 in order to achieve 99% collection efficiency.
The precipitator currently is designed for a gas flow of 512.48
m3/sec at 129°C (1,086,000 ACFM at 263°F), and the existing plate
area is 19,906 m2 (214,272 ft2). This gives a design specific col-
lecting area of 38.84 m2/(m3/sec) or 197 ft2/1000 ACFM. In view
of the expense required to enlarge the existing unit the required
amount, Iowa Public Service decided to evaluate the use of an S03
injection system for the purpose of lowering ash resistivity and
increasing the precipitator performance.
The SO3 injection system was supplied and operated by Research
Cottrell. The system burns molten sulfur to produce SO2 which is
subsequently oxidized to S03 in a catalytic reactor. The gas leav-
ing the converter is transported to the precipitator inlet duct
through an insulated line, and injection into the flue gas is accom-
plished with an insulated manifold. The system is designed with the
objective of maintaining the temperature in the transport line and
5
-------
in the manifold above the HaSCK dewpoint. The design details of
the SO3 system are not provided in this report as a result of a
confidentiality agreement with Research Cottrell.
-------
III. RESULTS FROM TEST PROGRAM
A. Particle Size Measurements
Particle size measurements were conducted during both the
baseline and S03 injection tests with cascade impactors and cyclones.
The purpose of the cascade impactor measurements was to provide
size distribution data for subsequent use in a mathematical model
which was employed to simulate the operation of the precipitator
during the test program. Multistage cyclones were used primarily
to obtain size fractionated samples for chemical analysis, and result:
from these measurements will be presented in the section on chemical
analyses.
During the baseline test series (March 27), a total of eight
Brink cascade impactor runs were performed at the precipitator in-
let, and two Andersen impactor runs were performed at the outlet.
Useful data were not obtained from the outlet runs because of sub-
strate sticking and stage overloading. Inlet impactors runs were
made with four-point traverses in which two points were approximately
0.3 meter (1 ft) apart at the top of the duct and the other two
points were the same distance apart at the center of the duct. The
traverses were conducted in port Nos. 2, 4, 5, and 7 (See Figure 20).
Figure 1 gives the data obtained from the March 27 inlet series on
log probability co-ordinates, and Figure 2 presents the distribution
in terms of cumulative mass loadings as a function of particle di-
ameter.
The second series of particle size measurements was conducted
the week of May 17, 1976. A total of sixteen Brink impactor runs
were conducted on May 18, 19, 20, and 21 using the same traversing
procedure described above. Outlet Andersen impactor runs were also
conducted on May 19, 20, and 21, but only the data obtained on May
19 are of interest because of problems encountered with the preci-
pitator or the injection system on the other test days. Figures 3
and 4 give the average inlet size distribution on log probability
-------
CO
H-
iQ 99.9
998
(D
* 99
2 H 98
0) 3
l"j 1^"^
s i s s g
(3ZISQ3I1
NOiinaiaisia i
et Size Distri
ch 27, 1976 wi
rap-rjuoo %Q6 *R
^inumo) uoT^nq
CUMULATIVE PERCEIV
(% LESS THAN IMDIC/
n o 8 8 S S
J rr u
O H-
CD <
CD 2
H
rf CD
"5 0.6
Q) 3 ,
0.2
r-1 rr
cn ^ 0.1
HI °-05
O
H
nni
IT- q~ffl |
-i _i. ......... 4... -fT^jj jjU^_^T4|i
~^^ -||plE" H
E===±:3±crp : ; ; Hi1, ^^ ^±r: ^jr ^r rr
r . , j _ i - 1 , i-j - 1 t T i T n iff-, i,--jnn Tt-r)n" j-«-^ ** . *-
-- "
^^^*=^=^Ms
M l.|i|| III
' W
=^=^^.'! | IMNJM^^,.
- i. .L.| I: :| 1 I 1 '
f> '
|-^ f-^r 3-^.-
rj! 1 1 -M t- -~i~-i i-H- 1 ' +-* -^ "rf T*** " "
4-. ' "' """'""""'jti't' tjjj iK^
1 1 j | -^> ]i>' p-j- » ' "" '"-lI'tT* *-*-"- 1 -
-=M=^=^=4=^^-=
i,; : , I, ,, , ;, ,,,! ...\\'. ,,
.<) , ,. «^
M i i ,,,'--,'', : ,,-,';,
j
-
r-
-
I
i
::-:S"£J|j5J
i i j ....]. -~~ f.;;,, .! r
V-
(III
I rff ' ffi 1
1 |||;
^44A|
^ I'l1!'! r *
if
||*-sg
i4fH4j-|
0.1
0.2
0.4 0.6 0.8 1.0 2. 4. 6.
PARTICLE DIAMETER ( micrometers )
8. 10.0
20.
40.
80. 80. 100.0
-------
I
Q
<
O
ERROR BARS INDICATE 90% CONFIDENCE INTERVAL
10'
10
10
10
m
i
E
i
'
ill
-ft
:
: ;
10
-1
10° 10
10
Figure 2.
PARTICLE DIAMETER ( micrometers )
Inlet Size Distribution (Cumulative Mass) for
March 27, 1976 with 90% Confidence Intervals
-------
99.99
-
UJ
Hi >
o <
S D)
^< (D
P- H
OD 3
I P-
H rt
tn
p- P-
^O N
-J CD
D
C P-
P- W
rt rt
P-
^> cr
o c
OP rt
P-
D O
O 3
3
i-ti .^
H- O
(D 3
O P-
(D (u
rt
M P-
3 <
rt fD
(D
< (C
DJ i^
P- O
w n>
3
rt
i
,-
G
-
.
4
I
~
_
_
E
0.01
0.6 0.8 1.0
2. 4. 6. 8. 10.0
PARTICLE DIAMETER I micrometers )
20.
40.
60. 80. 100.0
-------
10
ERROR BARS INDICATE 90% CONFIDENCE INTERVAL
5S!!!!!!!!!!!!!!!!!!!!!!!!!!«!!' ; ImSS!!!!"""""*"""" """""i
*»
i
10'
1
&
^:
1
II 1 IIIIII * tlMIIIII en i n j ' ,-
S iSiiiin ! i!! SS!
III! Illllll !* * IMIIIBBBI . I :-«! ^ - ;
IIIIIIIM ivmiiiiinnn .!. . , . .. '.:,** , .. .
:::: ifflliiiiiiiiiii::::: ::::::: :: ::::i:ii!!!iiiiii-i ::::!!!i
:::iSK;;iii!H!!Huuiii!;!itirniiH~~
10
1
PARTICLE DIAMETER ( micrometers )
Figure 4. Average Inlet Size Distribution (Cumulative Mass)
for May 18-21, 1976 with 90% Confidence Intervals
11
-------
co-ordinates and in terms of cumulative mass loading for the time
period May 18-21. Figures 5 and 6 are similar presentations of the
size distribution obtained on May 19 at the outlet of the precipi-
tator. The cumulative percent distribution is based on the total
outlet mass loading obtained with the impactor (35.45 mg/DNm3).
Because of the almost 10 meter (32.8 feet) depth of the outlet
duct work, full traverses were impractical. Therefore, single
point sampling was performed at a depth of 2.4 meters (7.8 feet).
This single point outlet size distribution, together with the
average of the size distributions obtained on May 18-21, were used
to compute the fractional collection efficiencies shown in Figure 7.
Note that a comparison of Figure 1 and 3 shows that essentially
the same inlet size distribution was obtained for the March and
May test series. The data from Figure 4 were used in the computer
model projections, which will be discussed in a subsequent section.
B. Mass Train Results
Inlet and outlet mass loadings for the baseline test were deter-
mined by Research Cottrell using the ASME Power Test Code 27 method
on March 27 and March 28. The inlet sampling location was upstream
of the injection manifold (Figure 20), and outlet data were obtained
in the stack. These data are given in Table 1.
Mass train data during the SO3 injection tests in May were
obtained by Kin Associates, Inc. using the following methods:
(1) A modified ASME sampling train at the inlet location with
an in-stack filter.
(2) A modified ASME sampling train at the stack location with
an in-stack Gelman type AE glass fiber filter.
(3) An EPA Method 5 sampling train at the stack. Four tests
were conducted with this equipment, but only one data set was con-
sidered useful due to problems with the precipitator or the S03
injection system. The tests were designated as follows:
12
-------
99.99
CJ
~
i
u
3 o
PJ q
>< rt
H
(-* (D
v£> rt
W
I-1 H-
U3 N
-~J (D
CTi
D
H-
CO
rt
H
H-
cr
:
O
c:
1
rt
H-
t
fD
M
'.
I
3
-
t-h
O
-
_
N
U
^
111
4
f.
:.
Ill
LL
--
0.01
0.1
0,2
0.4 0.6 0.8 1.0 2. 4. 6.
PARTICLE DIAMETER ( micrometers )
20.
60. 80. 100.0
-------
PARTICLE DIAMETER ( micrometers )
Figure 6. Outlet Size Distribution (Cumulative Mass) for
May 19, 1976
14
-------
Ul
0.01
0.05
0.1
nj 0.2
P-
1 1 1 Illl 1 Illl i i IIIM i inmiiii i i i iiiiniLi
IB ^ N ,00 g 0 §
O
NOII
gure 7. Fractional E
(Based on sir
i-" rn <
^5 H! £ 50
M p- £
' ' -
gEEE::::
:
""TIT TJT
1 i . , r
| ||
1.2
'
-nf-H-
-±ra
=Jt
.4
F*-
t=
"H4 ~~~
= ascpE
0.6
rrtr
o.e
j^-^p-
* T±-
jig l 1 ' M-J-I-
' ~~~ | "^^
1.0
GEOMETRIC ME
: , |i ^4||
- - ' ' ' * |-
^ 'I1!;1
^~~rr~r~i r
^Sffi ffiffi
j ,
4^44^
AN DIAMETEI
4i^LJ
P»
I" '"
4.
?
;|; ,, ; ,.r .,.
C«3^
FrTT
ll'll' 1 ' -^--
6. 8
MICROME1
_
j «^=
1
_^ j ; . , i , 1 . 1 ^
r==:E±:H:::l
. 10.0
ERS )
20.
^S^iic^^gl
S£S^^^^-
nft cii lip i-ll N + -^- 1%
''...' _.......
" i c
lllilliiill..ili.iill:i.lii.ll!li.l 'lil.
40. 60. 80.
}*1 99.8
| 95
=r 90
JS 70
I 60
£ 40
ffi 30
ft: 10
- 5
ij
OK
0.2
Iffi O-1
j~ 0.05
Ml o.oi
100.0
-.
-
-------
Table 1. Mass Train Results Reported by Research Cottrell
o\
Temperature,°C
Boiler Load (°F)
Date MW in Out
Gas Volume,
m3/sec
(ACFM)
In
Collecting1
Dust Concentration Area
gm/m3 (gr/ACF) Efficiency m2/(m3/sec)
3-27 299
3-282 299
107
(224)
111
(231)
96
(204)
93
(199)
(833
(740
417
,000)
349
,000)
585
(1,240,
386
(819,
000)
000)
6
(2
6
(2
.32
.76)
.04
.64)
\j
0.
(0
0.
(0.
'ui_ » |
482 92.4
.21)
593 90.2
259)
Itf/lUOO ACFMJ
34.02
(173)
51 .6
(262)
1. Based on outlet flow rates
2. Gas flow appears inconsistent with boiler load.
-------
Date Test No. Description
5/19/76 2-1 S02 converter temperature fell; S03
injection rate uncertain
5/19/76 2-2 Normal test
5/20/76 2-3 Interrupted due to TR set failure
5/21/76 2-4 Interrupted due to TR set failure
Table 2 summarizes the data obtained by Kin Associates on
May 19. The dissolved solids in the first impinger were obtained
to determine whether any S03 remaining in the gas phase at the stack
would appear in the residue after evaporation of the liquid. The
results in Table 2 indicate that the outlet mass loading is increased
about 18% if the impinger solids are included. Chemical analyses
of the impinger wash were also conducted, and these results are
discussed in a subsequent section.
It is of interest to compare the mass loadings obtained with
the mass trains in the stack with that obtained from the single
point measurement with the Andersen impactor at the precipitator
outlet. The Andersen impactor obtained a mass loading of 0.0227
gm/m3 (.00992 gr/ACF) which is only 32% of the results indicated
by the EPA train without inclusion of the impinger. Thus, the
fractional efficiencies plotted in Figure 7 are not representative
of the overall precipitator performance. It is probable that most
of the relatively large particles resulting from rapping reentrain-
ment are concentrated near the bottom of the duct and were, there-
fore, not captured by the impactor. The data in Figure 7 should,
however, provide a reasonably accurate representation of sub-two
micron collection efficiencies. Similarly, the total mass loadings
obtained with the limited impactor traverses at the inlet will show
significant disagreement with the mass train results, but the smal-
ler size fractions should be represented with sufficient accuracy
for use as input data to the precipitator computer model.
A comparison of the ASME-derived mass loadings and precipitator
efficiencies in Tables 1 and 2 shows that, for the brief period when
17
-------
Table 2. Mass Train Results Reported by Kin Associates
Gas
B . , Temperature, °c Volume
ouiier (°F) m /sec
Date MW m Out in AC out
5-19 303
129
(264)
122
(252)
557 476
(1,180,000) (1,009,000)
Dust Concentration
In
6.85
(2.99)
gm/m
ASME
0.0508
(0.0222)
' (gr/ACF)
Out
EPA & First
0.0703 0.0828
(0.0307) (0.0362)
Efficiency, %
EPA w/o
ASME Impinger
99.27 98.96
Specific
Collecting
Area
m2/(m3/sec)
(ft'/lOOQ ACFM)
41.82
(212)
Test
No.
2-2
-------
normal precipitator operation was possible, the precipitator per-
formance was increased to the extent that outlet mass concentrations
decreased by about a factor of ten as a result of the S03 injection.
This increase in efficiency is consistent with the improvement in
power supply performance and the decrease in dust resistivity, as
discussed below.
C. Resistivity Measurements
In situ resistivity data were obtained with a point-plane
probe1 during both the baseline and S03 injection test series. These
data are given in Table 3, and it is apparent that dust resistivity
during the S03 injection tests is about two decades lower than it
was during the baseline series. The data in Table 3 are obtained
with parallel plate cell geometry with an applied electric field
slightly lower than the value which is sufficient to cause spark-
over. It is also possible to obtain resistivity data from the
voltage-current characteristics of the apparatus with and without
dust on the collecting electrode as illustrated in Figure 8. The
collecting area of the measurement cell is 5 cm2, and resistivity
may be calculated from the dust layer thickness 0.09 cm, the volt-
age difference between the "clean" and "dirty" voltage-current
curves (5600 volts), a selected current (0.5 x 10~6A), and the cell
area. Thus,
n = (5600V) (5cm2) = 6>2 x 1Qn ohm-cm
p (0.5 x 10~6A)0.09 cm
at an applied field strength of 62.2 kV/cm with a current density
of 100 nA/cm2. In contrast, the parallel plate data were obtained
with an applied field of 16.6 kV/cm at a current density of 2 nA/cm
and the resistivity obtained under these conditions was 8.3x1012
ohm-cm. The resistivity value derived from the voltage-current
curves is expected to be lower than the parallel plate data under
these conditions as a result of electrical breakdown (back corona)
in the deposited dust layer.
19
2
-------
Table 3. In situ Resistivity from George Neal Plant, Unit 2
Baseline Test Series (3/27/76)
Resistivity
Temperature °C(°F) ohm-cm
110(230)
118(244)
118(244)
121(250)
8.3 x 1012
5.7 x 1012
5.9 x 1012
6.8 x 1012
S03 Injection Test Series
Temperature Resistivity
°C(°F) Date ohm-cm '
127(261) 5/17/76 4.4 x 1010
132(270) 5/19/76 3.1 x 10l°
138(280) 5/19/76 l.OxlO11
142(288) 5/19/76 4.1 x 1010
143(289) 5/20/76 3.5 x 1010
143(289) 5/20/76 4.1 x 10x°
20
-------
I ' I ' I
_ (5600X5.0) =C23t1011
Figure 8. Resistivity Probe Voltage-Current Characteristics
without SO3 Injection
21
-------
Figure 9 shows the resistivity probe voltage-current relation-
ships with SO3 injection. Note that the resistivity derived from
the voltage-current curve at 100 nA/cm2 is higher than the parallel
plate data because, with the relatively low dust resistivity, elec-
trical breakdown is not occurring in the deposited dust layer at
this current density. In general, the parallel plate data are con-
sidered more reliable than that derived from the voltage-current
curves.
D. Voltage-Current Characteristics of the Precipitator
Figure 10 illustrates the arrangement of the transformer-rect-
ifier sets on the precipitator. The power supplies are not equipped
with secondary voltage meters, and therefore voltage divider resis-
tors were attached to selected TR sets for the purpose of obtaining
secondary voltage readings. Figures 11 through 14 present the
secondary voltage-current relationships for the indicated TR sets
obtained during the baseline test on March 27. These curves in-
dicate that back corona and/or severe sparking occur at low values
of current density, which is indicative of high dust resistivity.
Note that the automatic operating point location is such that much
of the power input is not useful power for the precipitation process,
Figures 15 and 16 show the secondary voltage-current relationships
for TR sets 3 and 8 with S03 injection on May 21. The shape of
these curves, in contrast to Figures 11, 12, 13, and 14, indicates
that dust resistivity is not limiting the electrical operating condi-
tions. This conclusion is consistent with the in situ resistivity
measurements. However, as stated previously, TR sets were tripping
our during this test series due to dust removal problems. Average
electrical readings from the panel meters for March 27 and for May
19 are given in Tables 4 and 5, respectively.
22
-------
SO
Q_
0.5
1 I ' I ' T"7
>SPARK=4.1x1010
1400 (5)
(0.5x10-6)(0.1)
= 1.4x 10^ ohm-cm
10 12
kV
CLEAN
DIRTY
. I . I .
14
16 18
Figure 9. Resistivity Probe Voltage-Current Characteristics
with SO3 Injection
23
-------
t Gas Flow
CTD
CTD
r<
Figure 10. Arrangement of Transformer Rectifier Sets for
George Neal Unit 2
24
-------
0.6
0.5
0.4
| 0.3
0.2
0.1
0
I
O
- OA -
AVERAGE OPERATING^
POINT o
0
0
O
o
o
o
o
la I ' !
15 20 25 30 35 40
ou
13.4
10.8
o
<
H
8.1 1
Ul
Q
en
i»
CURRENT
2.7
kV
Figure 11. V-I Characteristics for TR Set 3 on March 27, 1976
25
-------
0.4
0.3
CO
&
0.2
0.1
~~i r~
AVERAGE OPERATING
POINT
_O_
15
20
25
30
kV
J L
10.8
8.1
5.4
2.7
E
u
Z
Ul
cc
cc
D
U
35
40
Figure 12. V-I Characteristics for TR Set 4 on March 27, 1976
26
-------
0.4
0.3
1 0.2
^
0.1
ft
AVERAGE OPERATING *-£
POINT
O
O
O
o
o
o
I
24.1 N
E
o
c
>."
t
16.0 |
UJ
Q
I-
Z
UJ
CC
cc
8.1 3
15 20 25 30 35 40
kV
Figure 13. V-I Characteristics for TR Set 7 on March 27, 1976
27
-------
to
Q.
U.O
0.4
0.3
0.2
0.1
0
1
I I I I
AVERAGE OPERATING ».* o
POINT U
o
O
0
0
0
o
0
I c I I I
40.2
32.2
M
O
24.1 1
1 .
P~
Z
UJ
Q
o>
b
CURREN1
8.1
5 20 25 30 35 40
kV
Figure 14. V-I Characteristics for TR Set 8 on March 27, 1976
28
-------
co
a.
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.
r- , . ,
_ CURRENT LIMIT *-
^^M
AVERAGE OPERATING »-£
POINT FOR 5/19/76
_
A ^^^~
*
I " '
0 " 20 30 40
kV
/.o
32.2
26.8
21.4
16.1
10.7
2.68
50
CM
I
V)
HI
o
I-
UJ
cc.
a.
u
Figure 15. V-I Characteristics for TR Set 3 on May 21, 1976
29
-------
Q.
<
I.O
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
11
i i i r-
CURRENT LIMIT
AVERAGE OPERATING-^
POINT FOR 5/19/76 ^^*^
~"~
,
_ _
-
0 ^HIBH
^^M* _ A
-
-
-
I - I
104.5
88.4
CM
O
64.3 <
I
V)
LU
48.2 Q
z
111
DC
QC
D
U
32.2
8.0
> 20 30 40 50
kV
Figure 16. V-I Characteristics for TR Set 8 on May 21, 1976
30
-------
TRtt1
7
8
4
3
Table 4. George Neal - Unit 2
Average voltages and currents for 3/27/76
Primary amps Primary volts Secondary amps
76
93
80
105
213
230
205
220
0.36
0.45
0.34
0.5
Current
Density
nA/cm2
29.0
36.2
9.2
13.4
1
2
5
6
40
30 to 40
140
145
110
170
150 to 200
280
275
225
0.07
0.06 to 0.14
0.8
0.82
0.43
1.8
( sparking,
meters
swinging)
64.3
65.9
no sparking
34.6
sparking
1 TRs 1, 2, 3, and 4 each have 40,176 ft2 collecting surface.
TRs 5, 6, 7, and 8 each have 13, 392 ft2 collecting surface,
for 214,272 ft2 total.
2. Average readings during sparking.
31
-------
Table 5. Average voltages and currents for 5/17/76
Current
TR No.
71
8
4
3
1
2
5
6
Primary Amps
86.1
191
199
164
137
220
195
193
Primary volts
215
346
295
328
306
312
353
349
Secondary
.474
1.156
1.137
.917
.728
1.397
1.234
1.248
"diO4_-£
amps nA/cm2
38.1
92.9
30.5
24.6
19.5
37.4
99.2
100.0
1. Declined from 1.18 amps to 0.15 amps at end of day.
32
-------
E. Chemical Analyses
1. Coal
Ultimate analyses of four coal samples (one collected during
the baseline test and three during the conditioning tests) are
given for the "as received" condition in Table 6. The data show
a reasonable degree of uniformity, as desired. The aspect of the
coal composition that is of primary interest is the sulfur concen-
tration, which is around 0.6% by weight.
2. Flue gases
The vapors of S02 and H2SOi» were determined with a sampling
train in which the H2SOif was first condensed around 70°C (160°F)
and the S02 was absorbed in a bubbler filled with aqueous H202.
Each sample was titrated as dilute HaSOi* with Ba(ClO^)2 and Thorin
as the endpoint indicator.2 The concentration of H20 vapor was
determined by condensing part as the liquid and absorbing the re-
mainder with silica gel. Concentrations of C02 and 02 were deter-
mined by Orsat analysis.
Theoretical concentrations of C02, HaO, Oa/ and S02 were com-
puted by using the coal analysis in Table 6 and assuming the com-
bustion air contained 2% of H20 vapor. The percentage of excess
air was not known; hence, predicted concentrations of the flue gases
were displayed in a graph as functions of excess air to permit a
comparison with the experimental results.
Table 7 gives the results of the experimental determinations,
and Figure 17 compares these results (except for H2S(K) with the
computed curves for varying percentages of excess air. The data
in the table indicate that during the conditioning tests the con-
centrations averaged 11.9% C02, 7.2% H20, 5.6% 02, and 431 ppm S02.
No reliable data were obtained for the first three of these gases
during the baseline test; however, the somewhat higher result for
SO2 during the baseline test indicates that a lower excess air
level was used during this test. The data for H2S04 were higher
33
-------
Table 6. Ultimate Analyses of Coal Samples
Moisture1
Carbon
Hydrogen2
Nitrogen
Sulfur
Ash
Oxygen2
Btu/lb
Baseline Test
3/27
6.86
60.83
4.88
1.26
0.66
10.45
15.06
10,949
Conditioning Tests
5/19N
6.28
62.24
5.56
0.65
0.57
10.66
14.05
11,047
5/19PM
5.91
59.72
4.60
0.88
0.61
13.84
14.44
10,579
5/21N Average
5.89 6.03
60.89 60.96
4.37
1.26
0.57
11.52
15.50
4.84
0.93
0.57
12.00
14.67
11,008 10,878
Determined by air drying.
n
Elements assumed to be present as chemically-bound water (all of
the oxygen and an equivalent amount of the hydrogen - less than
the total of hydrogen, of course).
34
-------
Table 7. Analyses of Flue Gas
Baseline
Test Conditioning Tests
Sampling Conditions
Date 3/27 5/19 5/19 5/20 5/20 5/21 5/21
Location Inlet Inlet Outlet Inlet Outlet Inlet Outlet
Temperature, °F 250 265 250 285 250 285 250
°C 121 129 121 141 121 141 121
Flue-gas Concentration
C02, % 11.7 11.6 11.2 11.0 12.9 12.9
02, % 6.0 6.0 6.3 6.2 4.6 4.5
H20, % 6.6 7.6 6.7 8.1 7.1 7.3
S02, ppm 521 473 419 446 391 428
H2SO^r ppm 0.4 1.2 0.9 10.6 0.6 8.0
35
-------
cc
o
8"
LL
O
3?
IU
5
D
O
16
12
10
C02
(C02)
H20
(H20)
1800
1600
1400
1200
1000
oo
800
600
400
200
EXCESS AIR,%
Figure 17. Comparison of Predicted Gas Concentrations
(Functions of Excess Air) with Observed
Concentrations (Displayed by Dashed
Horizontal Lines)
36
-------
during conditioning as expected; no average of these data is mean-
ingful, however, because of the variation in the sampling temperature
(a point later discussed in greater detail).
The comparison of observed and predicted concentrations in
Figure 17 shows the experimental results lying on the predicted
curves at excess air levels as follows: 02, 40% excess air; H2O,
>50% excess; C02, 42% excess; and S02, 38% excess. These comparisons
lead to two observations: (1) the experimental concentrations of
C02, 02, and S02 are consistent with the same excess air level
(about 40%) but (2) the concentration of H20 indicates a consider-
ably higher air level and is thus probably lower than the true
value (the estimated true concentration of H20 is about 8.5%) .
3. Fly ash
a. Overall oxide composition
Samples of fly ash were collected from selected hopper loca-
tions under the precipitator to be analyzed for overall composition
as expressed by oxide concentrations. The configuration of the hop-
per system is shown in Figure 18. The hoppers used in sampling
during the baseline test were Nos. 16 (inlet row) and 8 (outlet row);
those used in the conditioning tests were Nos. 17 (inlet) and 9
(outlet). All of the samples were thus taken from hoppers adjacent
to the midline of the precipitator.
For each inlet and outlet sample taken at a given time, a
composite was prepared to represent the appropriate amounts from
each source. The ratio of inlet sample to outlet sample in the
composite was 3.6:1.0 for the baseline test or 10.0:1.0 for the
conditioning tests.
The appropriate ratios were computed by assuming that for each
test mode (with conditioning or without) the effective migration
velocity w in the Deutsch equation was constant through the precip-
itator from inlet to outlet. In view of the fact that the electrode
area over the inlet hoppers is one-half of the total, equations in-
volving the total precipitation efficiency Et and the inlet
37
-------
GAS OUT
5
13
6
14
7
15
8
16
9
17
10
18
11
19
12
20
GAS IN
Figure 18. Hopper Configuration for the Precipitator
38
-------
precipitation efficiency E. can be written in terms of the respec-
tive electrode areas A and A. as follows:
ln(l - E ) = -Atw/V
ln(l - E±) = -A..W/V
= -0.5 Atw/V
Therefore,
= 0.5 ln(l - Et)
Ei = (1 - Et) °'5
E± = 1 - (1 - Et)°-5
The fractions of ash received by the inlet and outlet hoppers, res
pectively, are E± and (Efc - EI) . Hence, the ratio of masses is
given by
Inlet mass _ Ei _ 1 - (1 - Et)
Outlet mass ~ Et-E± Efc-l + (l-Et) ° 5
With Et taken as 0.924 for the baseline test, the computed mass
ratio is 3.6. Similarly, with Efc taken as 0.990 for the condi-
tioning tests, the result is 10.0.
A portion of each composite sample was ignited to determine
the weight loss during ignition and then it was divided into three
final fractions that were separately dissolved in (a) a mixture of
HF and H2SO,, (b) fused NaOH, and (c) fused Na2C03. The fraction
dissolved by acid was analyzed for Li, Na, K, Mg, Ca, Fe, and Ti
by atomic absorption spectroscopy and for P by a colorimetric pro-
cedure. The fraction dissolved in fused NaOH was analyzed color-
imetrically for Al and Si, and the fraction dissolved in fused car-
bonate was used for turbidimetric determination of S as SO 3.
The results of the analyses expressed as oxide weight per-
centages are given in Table 8. These data indicate that the major
difference (on a relative basis) in the samples from the baseline
and conditioning tests was in % S03; the difference was a gain of
about 0.67% as the result of S03 injection during the conditioning
tests.
39
-------
Table 8. Analyses of Hopper Ash
Weight %, Baseline Test
Weight %, Conditioning Tests
Component
Li20
Na2O
K20
MgO
CaO
Fe203
A1203
SiO2
Ti02
PiOs
S03
LOI
3/27AM
0.
0.
1.
2.
13.
6.
17.
53.
1.
0.
0.
0.
01
39
9
6
6
7
9
0
3
2
78
39
3/27AM
0.01
0.34
1.9
3.0
15.0
6.5
19.6
48.7
1.0
1.1
0.75
0.18
3/27PM
0.01
0.34
2.0
2.3
14.4
6.4
19.5
51.9
0.9
0.9
0.75
0.21
Avg.
0.01
0.36
1. 9
2.7
14. 3
6.5
19.0
51.1
1.1
0.7
0.76
0.33
5/19AM
0.
0.
1.
2.
15.
5.
17.
50.
0.
0.
1.
0.
01
45
9
9
0
7
9
3
9
7
38
51
5/19PM
0.01
0.45
2.0
2.9
13.1
5.9
17.9
51.7
0.9
0.7
1.66
0.58
5/20
0.01
0.39
1.8
2.4
13.2
5.9
17.2
53.0
0.8
0.6
1.26
0.50
5/20
0.01
0.46
1.7
3.0
14.6
6.0
18. 3
53.4
1.1
0.7
1.42
0.44
Avg.
0.01
0.44
1.9
2.8
14.0
5.9
17.8
52.1
0.9
0.7
1.43
0.51
-------
b. pH and soluble sulfate concentration
Determinations of the pH values of fly-ash slurries in dis-
tilled water and the concentrations of SO.*"2 dissolved in the
slurries were made for ash samples from several sources: (1) pre-
cipitator hoppers, (2) cyclones used for sampling from the flue-gas
ducts, and (3) filters used for sampling from the stack. For the
determination of pH and soluble SOi*"2 values, 0.1 g of ash was
mixed with 30 ml of distilled water and stirred until the pH
reached a stable value. This pH value was then recorded; the
liquid phase was separated from the suspended solids and analyzed
for S04~2
(1) Hopper samples. The data for hopper samples are given
in Table 9. The pH values listed are in the highly alkaline
range; all are above pH 11 and show no significant variation with
hopper source or with sampling conditions (with or with S03 in-
jection) . Within the first minute or so after addition of fly
ash to water, however, the samples taken during S03 injection
showed evidence of free H2SO.t on the ash surfaces. Minima in the
range pH 4-5 occurred with the conditioned samples, but such pro-
nounced minima with unconditioned samples were not usually observed.
The eventual rise in pH to values above 11 is attributed to the
excess of soluble base toward the interior of the ash particles.
The data for SCK"2 in ash samples taken across the inlet to
the precipitator are plotted in Figure 19. Across the bottom of
the horizontal axis, hopper numbers are shown to identify the
locations within the precipitator where the samples originated;
across the top, temperatures measured in the inlet duct upstream
from the hoppers are given to show the temperature gradient result-
ing from the Ljungstrom air preheater. The data for two complete
sets of hopper samples taken during S03 injection are plotted and
connected by line segments; averages of results for individual
hopper samples taken with and without injection are also plotted.
41
-------
Table 9. pH Values and Soluble SO.,'2 Concentrations1 of Hopper Ash
Baseline Test Conditioning Tests
Hopper No. 3/27 AM 3/27 PM 5/19 N 5/19 PM 5/20 5/21
Inlet Outlet pjl SO.."2 pj_ SO.T2 pJJ SOi."2 PJL SO.."2 pj^ SO.."2 pH SOt
to
13 11.7 0.7
14 11.5 1.6
15 11.5 1.3
11.3 2.0
11.3 1.6
7 11.4 1.7
16 11.6 0.7 11.5 0.3 11.7 0.7
8 11.7 0.5 11.7 0.5 11.4 1.4
17 11.6 1.1 11.5 1.2 11.7 0.4
9 11.6 1.3 11.5 1.3 11.5 i.o
18 11.6 1.3
10 11.6 0.7
19 11.6 1.1
H 11.6 0.7
20 11.6 0.7
12 11.6 0.7
1Weight percentage
-------
3?
N
it
O
GO
UJ
_J
CQ
O
CO
GAS TEMPERATURE
°C 138
°F 281
139
283
136
276
129
265
123
254
121
249
118
244
116
241
2.0
1.6
1.:
0.8
0.4
a
o
A
3/27
5/19
5/20
5/21
13
14
15
16 17
HOPPER NUMBER
18
19
Figure 19. Soluble SOiT2 in Hopper Ash as a Function of
Hopper Location or Gas Temperature
43
-------
The variation in SO, 2 across the precipitator inlet shows an
essentially continuous, downward trend with decreasing gas tempera-
ture during SO3 injection on May 21, but show no consistent trend
during injection on May 20. A possible explanation of either vari-
ation is a lack of uniformity in the rate of flow of S03 into the
inlet gas duct. Another possible explanation for the more or less
regular trend on May 21 is the effect of temperature. However, the
observed direction of the trend (simultaneous decreases in both SO^-2
and temperature) is opposite to that expected from previous data in
another study of S03 conditioning3 or expected from the observed H2so
concentrations in the gas phase at different temperatures (Table 7).
(2) Cyclone samples. Locations in the gas ducts where the
series cyclones were used for sampling fly ash during conditioning
tests are indicated in Figure 20. The numbers circled in this
diagram are subsequently used to identify the different samples.
It may be seen from the diagram that two samples were taken upstream
from the line of S03 injection nozzles on opposite sides of the
duct at estimated temperatures of 138°C (280°F) and 115°C (240°F),
another three samples were taken between the nozzles and the pre-
cipitator (again at different temperature extremes), and a final
set of three samples were taken at one location in the outlet duct
near the stack.
The results of determinations of pH and soluble SO.T2 are
given in Table 10. Data are given for each size fraction and for
weighted composites of the inlet samples, but only for composites
of the outlet samples (which were of such limited quantity to pre-
vent study of each fraction).
The pH data, in general, show increasing acidity with decreas-
ing particle size or with increasing available S03 as the result of
injection. The SO.,'2 data show the same effects.
The only location in the flue-gas train where cyclone samples
were collected during the baseline test was at the outlet of the
precipitator. A composite of different size ranges of this sample
gave these results: pH = 10.8; % SO^-2 = 2.6.
44
-------
PRECIPITATOR
t
GAS FLOW
S03 INJECTION NOZZLES
2345678
PORT NUMBER
Figure 20. Locations for Sampling with Series Cyclones
45
-------
Table 10. pH Values and Soluble Sulfate
Concentrations of Cyclone Samples of Ash
Sample
3a
3b
5a
5b
5c
Source
Inlet before
injection
Inlet before
injection
Inlet after
injection
Inlet after
injection
Inlet after
injection
Outlet
Outlet
Outlet
Relative Size
Temperature1 Fraction2
Soluble
pH
High
Low
High
High
Low
C
M
F
Comp.
C
M
F
Comp.
C
M
F
Comp.
C
M
F
Comp.
C
M
F
Comp.
Comp.
Comp.
Comp.
11.2
9.8
8.7
11.2
9.6
8.7
11.1
11.1
10.9
11.3
9.5
6.7
11.1
11.2
10.8
7.8
8.2
8.7
0.27
0.92
1.5
0.30
0.34
0.84
1.6
0.37
1.2
2.0
3.7
1.6
0.82
3.4
6.6
1.1
1.2
1.4
2.8
T74
6.4
6.3
7.6
'High temperature, ca 138°C (280°F), Low Temperature, ca 115°C (240°F)
2C, M, and F correspond to DBO values of 2.2, 0.8, and 0.5
um, respectively. Comp. indicates a composite of all size
ranges.
46
-------
(3) Filter samples. Samples of the fly ash collected on the
filters that were used at the precipitator outlet (in the stack)
during the efficiency tests were analyzed. These samples were from
both the ASME and EPA sampling trains. The results are given in
Table 11.
4. Gases absorbed in the EPA mass-sampling train
Analyses were conducted to determine whether nonvolatile
material was collected in the impingers of the EPA mass-sampling
train at the outlet of the precipitator. Weights of solid material
in the first impinger catches were taken into account in calculating
the mass concentrations. These weights were determined by drying
aliquots of the material caught by use of a temperature high enough
to evaporate water rapidly. However, a less vigorous drying pro-
cedure could leave a residue of absorbed gases-specifically includ-
ing H2SCU absorbed from the filtered flue gas or H2SCH produced by
absorption of S02 from the flue gas and oxidation of the S02 in
the absorption medium.
In view of this possibility, we performed analyses for absorbed
gases in aliquots of the first impinger catches and in composites
of the second and third impinger catches. One analytical method
employed with a few samples was to titrate absorbed acid with NaOH;
the results indicate that comparable amounts of two acids were
present: HzSOi, and H2SO3 (sulfurous acid, or absorbed but un-
oxidized S02). Reasoning that the H2SO3 would be lost during any
drying procedure, we then removed the H2S03 and determined the
remaining quantity of SO,'2 (not only the H2SCH evidently present
but also any SCK"2 leached from solid material) .
The results of the SOiT2 determinations indicated that vir-
tually equal quantities appeared in the first impinger catch and
in the composite of the second and third impinger catches. The
total amount of SO,'2 found was equivalent to about 5 ppm of HzSO*
vapor, which could give a particulate concentration of about
.02 gm/m3 (0.01 gr/CF) * as H2SOi, mist in the plume. As such, the
* Concentration at 21°C and 1 atm.
47
-------
Table 11. pH Values and Soluble SO,, 2
Concentrations of Filter Samples of Ash
Test
2-1
2-2
2-3
2-4
Sampling
Train
AS ME
EPA
AS ME
EPA
AS ME
EPA
AS ME
EPA
pj
10.
10.
9.
9.
10.
9.
[
4
4
9
6
5
_ i
5
_ i
Soluble
5.1
4.9
5.6
3.4
4.2
i
6.2
i
aFilter sample not available for analyses.
48
-------
SOi*"2 found would be significant. However, the finding of S0i*~2
concentrations in the second and third impingers that were nearly
equivalent to those in the first impinger suggests that much of
the SO.*"2 found was from absorbed and oxidized S02 , and not from
H2SCU vapor. As S02 in the flue gas, the material found would not
be logically included in the particulate emission.
49
-------
IV. MASS BALANCE CONSIDERATIONS
A. Sulfur Excluding Injected SOa (Conditioning Test)
It is possible to use the analytical data given in Section III.E
to compare the rate of combustion of sulfur in the coal with the rates
of production of sulfur compounds in the flue gas and in the fly ash.
We made this comparison by using the data from the conditioning tests
with carbon as a basis for comparison. In other words, we assumed
that all of the carbon in the coal appeared as C02 in the flue gas
and then compared the mole ratio to sulfur to carbon in the fuel
against the mole ratio in the combination of flue gas and fly ash.
For the fuel, the average weight ratio of sulfur to carbon is
0.57/60.96 (Table 6), which corresponds to a mole ratio of:
0.57/32.07 _ 3 c ln-3
60.96/12.01 ~ U
For the flue gas, the average concentrations of SOa and COa
are approximately 430 ppm and 12% by volume, respectively (Table 7) .
The concentration of SOa in the gas phase prior to injection of this
compound is negligible (less than 1 ppm). However, the concentration
of SO 3 in the fly ash is not small enough to be ignored. If the
weight percentage of S03 in the ash prior to treatment with this gas
from the conditioning system is taken as 0.76% (Table 8) and the fly
ash concentration entering the precipitator is computed on an abso-
lute basis (that is, for moist flue gas) as 9.00 mg/m3,* then the
SO3 in the ash corresponds to a gas-phase composition of:
0.0076 x 9.00 gm/m3 n oc .rt_3 . . 3
- 80 * = 0.85 x 10 mol/m3
This concentration is added to the concentration of SO^, which is
430 x 10"6 m3/m3 _ ,_ , -3
24.1 x 10-3 ~ 17<8 X 10
* This concentration represents the average of several inlet
determinations, expressed for a temperature of 21°C at a
pressure of 1 atm with water vapor present.
50
-------
The concentration of COz is:
0.12 mVm3 = 5 0 mol/m
b-U mol/m
24.1 x 10-3 m'/mol
Thus, the mole ratio of sulfur to carbon in the combination of
flue gas and fly ash is given by:
(17.8 + 0.85) x 10"3 _ o -7 v irr3
- 57o 3'7 x 1U
Three conclusions are now evident. First, the sulfur/carbon
ratio in the combustion products (3.7 x 10" 3) agrees remarkably
well with the ratio in the fuel (3.5 x 10~3). Second, there is
evidently little of the sulfur in the fuel that is discharged from
the boiler as bottom ash. Third, if it is assumed that all of the
sulfur in the fuel is initially oxidized to S02 but that subsequent
partial oxidation of S02 to S03 occurs in the flue-gas train as
the temperature is lowered, then the conversion factor of S02 to
SO 3 is best represented by the concentration of S03 in the fly ash
(not in the flue gas). The computed conversion factor is:
_ °-85 x 1?"3 1n-TX 100 = 4.4%
(17.8 + 0.85) x 10-3
B. Injected SO 3: Quantity Accounted for at the Inlet of the
Precipitator
1. S03 present as H2SCH vapor in the gas phase
Data on inlet concentrations of H2SCK vapor are necessarily
limited to the results that could be obtained at the only two sam-
pling ports available between the injection nozzles and the pre-
cipitator. The data (given previously in Table 7) were for dif-
ferent temperatures, around 130°C (265°F) toward one side of the
duct and around 143 °C (290°F) on the other side. They show higher
concentrations of H2SCH at the higher gas temperature, as expected.
To obtain a reasonable approximation of the average H2SCU con-
centration at the precipitator inlet, the following approach was
51
-------
taken: First, Figure 21 was prepared to compare the H2SOi, concen-
trations found at different sampling temperatures with the concen-
trations predicted from the data of Banchero and Verhoff1* at the
dew point (theoretically, the maximum vapor concentrations attain-
able) . As expected, the experimental results are portrayed by a
curve falling substantially below the dew point curve. Next, Figure
22 was prepared to show average gas temperatures at different dis-
tances across the inlet duct (these temperatures were computed from
the data of Kin Associates that were obtained in pitot traverses).
Also plotted in Figure 22 were H2SO» concentrations at these tem-
peratures as obtained by interpolation of the previous curve
(Figure 21) for the experimental values. Finally, the average H2SCH
concentration across the duct was computed from the relationship
between apparent experimental concentration and location in the
duct.
The value of the average thus obtained is 1.8 ppm. It repre-
sents the approximate increase in the H2SOif vapor concentration as
the result of S03 injection (the concentration during the baseline
test has been reported as 0.4 ppm but is not large enough to be
clearly distinguished from zero).
2- SO3 present as SCK"2 in the fly ash
Data based on hopper samples. Average values for the total
percentage of S03 in fly ash collected in hoppers adjacent to the
midline of the precipitator are 0.76% without conditioning and
1.43% with conditioning (Table 8). Using the difference of
0.67% as a measure of the effect of S03 injection and taking the
value of 9.00 gm/m3 as representative of the fly-ash concentration
at the inlet of the precipitator,* one calculates as follows to
obtain the concentration of injected S03 thus accounted for:
* This concentration represents the average of several inlet
determinations, expressed for a temperature of 21°C at a
pressure of 1 atm with water vapor.
52
-------
a
a
*
O
X*
16
14
12
10
8
110
230
116
240
121
TEMP °C
127 132
138
143
149
_L
T
T
DEW POINT CURVE
(8.5% H20)
°/
/
/
/ EXPERIMENTAL
/ RESULTS
»o
w
250
260 270
TEMP°F
280
290
300
* u cr» r-nnr-^nf-rations at the Dew Point
Figure 21. Comparison of H2SOi» Concenrratiuaa
with Experimental Results
53
-------
160
320
104
Figure 22. Relationship between Flue-Gas Temperature and
Experimental H2SO^ Concentrations
54
-------
0.0067 x 9.00 gm/m3 = 1Q_3 mol/m3
80 gm/mol
0.75 x 10~3 mol/m3 x 24.1 x 10~3 m3/mol =18.0 x 10'6 m3/m3
In other words, VL8 ppm or about 75% of the 25 ppm of injected S03
is accounted for. In connection with this result, it is necessary
to point out that the composition of fly ash collected near the
middle of the precipitator may not be representative of the total
collected, particularly in view of the erratic nature of results
for soluble SO.,"2 in ash across the entire inlet of the precipi-
tator (Figure 19).
Data based on cyclone samples. Results of determinations of
soluble SO,"2 in fly ash collected in cyclones during S03 injection
give another basis for calculating the fraction of the conditioning
agent found in the fly ash. Two composite samples taken upstream
from the injection nozzles were found to contain an average of
about 0.34% soluble SO,'2 (Table 10). Corresponding samples taken
between the nozzles and the precipitator contained individually
1.6, 1.1, and 1.4% soluble S04"2 or averaged 1.37%. The difference
in the averages, 1.03%, upstream and downstream of the nozzles
corresponds to the following concentration of injected S03:
0.0103 x 9.00 gm/m3 x 24 L x 10-s m3/moi = 23 x 10~6 m3/m3
96 gm/mol
The result, 23 ppm, is very close to the nominal value of the in-
jected concentration.
C. Injected SQ3; Quantity Accounted for in Stack Emissions
1. Emission as H2SQi» vapor
Concentrations of HZSO» found in the outlet duct during con-
ditioning tests were all less than 1 ppm. Such low values are to
be expected as a result of the low gas temperature recorded at the
sampling point 120'C (about 250°F), if one consults the dew point
concentrations predicted by Banchero and Verhoff (Figure 21). It
is therefore evident that if S03 injection significantly increases
55
-------
SOi, 2 emissions, the increase must be found in the composition
of emitted particulates.
2. Emission of S0it~2 in particulates
Data from cyclone samples. The soluble SO^2 concentrations
in particulate collected in cyclones at the outlet duct during con-
ditioning tests were individually 6.4, 6.3, and 7.6% or the average
was 6.8% (Table 10). The average value was used with outlet parti-
culate concentrations calculated from efficiency tests to compute
equivalent concentrations of S03 lost to the stack in the solid
phase as summarized in Table 12.
The soluble S0i»~2 content of outlet ash during the baseline
test was 2.6%. The particulate concentration at this time was about
0.63 mg/1.* Thus, the corresponding S03 concentration without con-
ditioning was 4.4 ppm, a value that exceeded each emission level
given in Table 12 except those for Test 2-3, which is known to have
given a faulty indication of the precipitator performance during
conditioning. It is evident, therefore, that the total amount of
SO3 emitted in particulates was lower with conditioning than without
as a result of the marked reduction in the mass concentration of
particulates.
Data from filter samples. Computations of stack losses of
SO3 as particulate SOi,"2 found on the filters used in the efficiency
tests gave the results summarized in Table 13. These results con-
firm the conclusion just given: an insignificant fraction of the
injected S03 escapes to the stack as SOi,"2 in particulates.
* This concentration is the average of two outlet determinations,
expressed for a temperature of 21°C at a pressure of 1 atm, with
water vapor present.
56
-------
Table 12. Injected S03 Found
In Outlet Cyclone Samples of Ash
Sampling Particulate Con- Corresponding S03
Test Train1 centration2. gm/m3 Concentration, ppm
2-1 ASME 0.092 1.6
EPA 0.163 2.8
2-2 ASME 0.071 1.2
EPA 0.098 1.7
2-3 ASME 0.318 5.4
EPA 0.556 9.5
2-4 ASME 0.205 3.5
EPA 0.210 3.6
1. Used for determination of total particulate
concentration given in next column.
2. At 21°C and 1 atm. with water vapor present,
57
-------
Table 13. Injected SO3 Found
In Outlet Filter Samples of Ash
Sampling Particulate Con- Corresponding S03
Test Train centration1, gm/m3 Concentration/ ppm
2-1 ASME 0.092 1.2
EPA 0.163 2.0
2-2 ASME 0-071 1.0
EPA 0.098 0.8
2-3 ASME 0.318 3.3
EPA 0.556 2
2-4 ASME 0.205 3.3
EPA 0.210 z
1. At 21°C and 1 atiru with H20 present.
2. Filter sample not available for analysis.
58
-------
V. COMPUTER MODEL PROJECTIONS OF PRECIPITATOR PERFORMANCE
The SRI-EPA mathematical model6 was used to analyze the perform-
ance of the precipitator during the test program and to estimate
the plate area required for 99% collection efficiency without the
aid of conditioning agent. In order to use the model to simulate
the precipitator operation with high dust resistivity, it is neces-
sary to estimate a "useful" input power from the voltage-current
relationships. This is required because the model is based on the
assumption that the input values of current and voltage represent
a unipolar particle charging process in which all of the charge
transported to the collecting electrode from the corona wires is
carried by either corona current or the charged dust particles.
The average conditions used were estimated from the voltage-current
relationships obtained on March 27. The averages of the values
selected for the TR sets are: 35kV applied voltage, 3.5 nA/cm2
current density. Figure 23 presents results from the computer pro-
gram in terms of overall collection efficiency as a function of
specific collecting area. The theoretical performance is clearly
much greater than the measured performance obtained by Research
Cottrell on March 27. However, if the overall collection efficiency
is reduced by empirical relationships which are intended to esti-
mate the effects of gas sneakage, particle reentrainment, and non-
uniform gas velocity distribution, fair agreement can be obtained
by assuming a gas velocity distribution with a normalized standard
deviation of 25%, and by further assuming that reentrainment and
gas sneakage losses amount to 20% of the mass collected in each
stage over three effective stages. The computer projection based
on these assumptions is labeled curve 2 on Figure 23. These assump-
tions indicate that a specific collecting area of 78.8 m2/(m3/sec)
(400 ftVlOOO ACFM) would result in 99.0% overall collection effic-
iency. However, this projection does not contain a safety factor
for TR set failures, nor does it consider the possibility that the
59
-------
s?
u
LU
o
il
o
HI
o
u
99.9
Figure 23.
FROM RESEARCH COTTRELL TEST OF 3/27
700 ft2/1000ACFM
138 m2/(m3/sec)
Pr°jections of Collection-Efficiency
60
-------
assumed electrical operating conditions may be degraded if dust
resistivity increases. The latter consideration constitutes a
major uncertainty since the electrical operating characteristics
of the precipitator will be critically influenced by dust resis-
tivity changes.
In an effort to obtain an estimate of specific collecting
area which includes a safety margin, the following procedure was
employed to obtain curve 3 on Figure 23: (1) The model was used
to calculate the overall efficiency if the SCA were effectively
reduced from 78.8 to 59.0 m2/(m3/sec) (400 to 350 ft2/1000 ACFM) in
1/2 of the precipitator. (2) The model was used to calculate over-
all efficiency if the SCA were reduced from 118 to 88 m2/(m3/sec)
(600 to 450 ft2/1000 ACFM) in 1/2 of the precipitator. (3) The
results from (1) were plotted at 78.8 m2/(m3/sec) {400 ft2/1000 ACFM),
and the results from (2) were plotted at 118 m2/(ra3/sec) (600 ft2/1000
ACFM). This procedure results in an estimated requirement of 108
m2/(m3/sec), or 550 ft2/1000 ACFM, to achieve 99% collection effi-
ciency with a safety margin for transformer-rectifier failures as
specified above.
Figure 24 gives the theoretical model projections and the
reduced projections (using the same parameters for gas velocity
distribution and reentrainment and sneakage as were used in Figure
23) with the average electrical operating conditions achieved with
S03 conditioning. These results indicate that, with the measured
size distribution and electrical operating parameters achieved dur-
ing the SO3 injection test, 99% collection efficiency is theoret-
ically possible at the existing SCA. However, as stated earlier,
additional testing is required to determine whether the S03 system
will allow precipitator performance to remain at the 99% level for
extended periods with a boiler load of 300 MW.
61
-------
99.99
(J
z
LLJ
O
o
O
u
99.90
99,5
99.0
95.0
ASME ON TEST 2-2
EPA ON TEST 2-2
a = 0.25
S = 0.2
N = 3.0
I
100 200 300 400 500 600 ft2/1000 ACFM
19.7 39.4 59.1 78.7 98.4 118 m2/(m3/sec)
Figure 24. Computer Model Projections of Collection-Efficiency
with SO3
62
-------
ACKNOWLEDGEMENTS
Appreciation is expressed to Research Cottrell and Kin
Associates, for assistance in obtaining mass loading measurements,
and to Iowa Public Service Company for financial and technical
assistance. Mr. G. H. Marchant, Jr., supervised the Institute
personnel in the field, who consisted of several members of the
Physics Section and the Physical Chemistry Section.
Submitted by:
Edward B. Dismukes,
Senior Research Advisor
John P. Gooch, Head
Control Device Research Division
63
-------
REFERENCES
Grady B. Nichols, "Techniques for Measuring Ply Ash Heals-
tivity", EPA-650/2-74-079, August 1974.
J. S. Fritz and S. S. Yamamura, Anal. Chem. 27, 9
(1955). "
E. B. Dismukes, "Conditioning of Fly Ash with Sulfur
Trioxide and Ammonia", EPA-600/2-75-015, August 1975.
J. T. Banchero and F. H. Verhoff, J. Inst. Fuel 48, 76
(1975).
64
-------
TECHNICAL REPORT DATA .
(Please read Inuructivns on the reverse before completing!
1. REPORT NO.
E PA- 600/2-
77-242
4. TITLE AND SUBTITLE
Fly Ash Conditioning with Sulfur Trioxide
December 1977
PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Edward B. Dismukes and John P. Gooch
. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
1AB012; ROAP 21ADL-027
11. CONTRACT/GRANT NO.
68-02-2114, Task6
12. 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 PER
Task Final; 3-11/76
14. SPONSORING AGENCY CODE
EPA/600/13
T5. SUPPLEMENTARY NOTES IERL_RXP project officer is Leslie E. Sparks, Mail Drop 61, 919/
541-2925.
l.
bi
i6. ABSTRACT The report describes an evaluation of an SO3 injection system for the Georg
Neal Unit 2 boiler of the Iowa Public Service Co. in Sioux City, Iowa. Results of base
line tests without conditioning indicate a dust resistivity of 6 x 10 to the 12th power
ohm-cm at 118 C: the precipitator's average collection efficiency was 91.3% at a
specific collecting area of 42.8 sq m/(cu m/sec). Because transformer-rectifier sets
tripped out, apparently due to ash buildup in the hoppers, only one precipitator effi-
ciency test was conducted with the SOS system operating continuously with all T-R
sets operating. Results of this test were: (1) specific collecting area = 41.8 sq m/(cu
m/sec); (2) collection efficiencies = 99.27% (ASME method), 98.96% (EPA method),
and 98.78% (EPA method, including first impinger residue); and (3) 4x 10 to the 10th
power ohm-cm dust resistivity at 143 C. An adequate accounting was made for the fate
of the injected SOS.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
. COSATI Field/Group
Air Pollution
|Fly Ash
Treatment
Sulfur Trioxide
Dust
Electrostatic
itators
Electrical Resis-
tivity
Air Pollution Control
Stationary Sources
13B
2 IB
07B
11G
20C
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS
Unclassified
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
65
------- |