Baghouse

DOE/MC/lO637-3312
(DE93000252)
_miou Ca_sory UC-109
".
DOE/MC/10637--3312
DE93 000252
'.
• Pulse-Jet Baghouse
Performance Improvement
With Flue Gas Conditioning
Topical Report
Stanley J. Miller
Dennis L. Laudal
Work Performed Under Contract No.: DE.FC21-86MC10637
For
U.S. Department of Energy
Office of Fossil Energy
Morgantown Energy Technology Center
P.O. "Box880
Morgantown, West Virginia 26507-0880
By
University of North Dakota
Energy and Environmental Research Center
Box 8213
University Station
Grand Forks, North Dakota 58202-8213
October 1992
ACKNO_GMENT
The authors wish to acknowledge that the work reported in this document was jointly
funded by the U.S. Department of Energy (DOE), the Electric Power Research Institute
• (EPRI), and the Canadian Electrical Association (CEA). We thank Dr. Perry Bergman of
DOE PETC, Dr. Ramsay Chang of EPRI, and DeLip Deshpande of Alberta Power for their
helpful technical input throughout the study.
The authors also wish to thank TU Electric for supplying the Big Brown coal, Alberta
Power for supplying the Milner coal, and the ARCO Coal Company for supplying the Black
Thunder coal.
ABSTRACT
A pilot study was conducted at the Energy and EnvironmentalResearch Center

(EERC) at the University of North Dakota to evaluatethe effectiveness of flue gas
. conditioning in reducing tube sheet pressure drop and fine particulate emissions from a pulse-
jet fabric filter. The project was jointly funded by the U.S. Department of Energy (DOE),
the Electric Power Research Institute (EPRI), and the Canadian _ectrical Association
(CEA). The work was completed with EERC facilities consisting of a pulverized coal-fired
combustor and pilot baghouse. FuU-scalepulse-jet bags were employed under conditions
similar to large-scale baghouses. The investigation included baseline iests and tests in which
ammonia and SO3 were injected upstream of the baghouse to determine the effect of
conditioning on baghouse performance. The primary independent variables included coal
type, conditioning agent concentrations, air-to-cloth (A/C) ratio, and fabric type. The main
dependent variables were particulate emissions, baghouse pressure drop, and cohesive
properties of the fly ash. Results demonstrated significant benefits of using conditioning with
a pulse-jet baghouse, including a substantial reduction in particulate emissions and a
substantial reduction in pressure drop (or the ability to operate at a higher A/C ratio without
increasing pressure drop or bag-cleaning frequency). The improvements in fabric filter
performance correlatestrongly with a shift in the tensile strengthand with increases in the
aerated and packed porosity of the fly ash. Conditioningappearsto be applicable to a wide
range of coals and fabrics. Applications for this technology are where there is a need to
reduce pressure drop and/or particulateemissions in existing baghouses, to reduce fine-
particle air toxic emissions which may be required in the future, and for new baghouse
installations to allow operation at a higher A/C ratio while providing an ultrahigh fmc-
particle collection efficiency.
TABLE OF CONTENTS
ii
LIST OF FIGURF_ ...........................................
LIST OF TABLES ..................... ix
• 1.0 PROJECT SUMMARY ..................................... l
2.0 PROJECT OBJECTIVES .................................... 4

5
3.0 SCOPE OF WORK .......................................
4.0 FACIIXrY DF,,SCRIFFION AND PROCEDURF.S .................... 99

4.1 Particulate Test Combustor ..............................
4.2 Particulate Sampling and Measurement ........................ l0
4.3 Sulfur Trioxide and Ammonia Injection ....................... 15
4.4 Coal and Fly Ash Analyses ............................... 15
..................... 17
5.0. RF..SULTS ........................ 17
5.1 Coal Description and Analyses .............................
5.2 Theoretical Considerations ............... 20
5.3 Task 2 Results .................. 22
5.3.1 Big Brown Tests ................ 22
. • • • • , • • • • • • • • • • • •
5.3.2 Milner Tests ................. 37
5.3.3 Cohesive Properties and Ash Analyses .................... 42
5.4 Task 3 Results ...................................... 50
5.4.1 100-hour Tests with Big Brown Coal .................... 50
5.4.2 100-hour Tests with Milner Cc_ ....................... 59
5.4.3 100-hour Tests with Black Thunder Coal .................. 65
5.4.4 16-ft/rain Tests with l_ig Browu Coal .................... 72
5.4.5 Cohesive Properties and Ash Analyses .................... 81
81
5.4.5.1 Particle Size ................. 84
5.4.5.2 Tensile Strength ................................
5.4.5.3 Porosity Measurements ........................... 91
5.4.5.4 Ammonia and Sulfate Analyses ....................... 94
5.4.5.5 Residual Dust Cake Weight ......................... 98
6.0 CONCLUSIONS AND DISCUSSION ............................ 99

6, I 8-hour Tests with Milner and Big Brown Coals .................. 99
" 6.2 100-hour Tests with Big Brown, Milner, and Black Thunder Coals ..... 110
6.3 16-ft/rain Tests with Big Brown Coal ........................ 111
6.4 Unresolved Issues and Areas for Future Research ................ 111
113
7.0 REFERENCES .........................................
LIST OF FIGURES
Fibre
1 Combustion system used for tests, showing the PTC and baghouse ........ 10
2 Schematic of aerodynamicparticulatesizer........................ 12
3 Schematic of the paxticulate-samplingsystem (APS 33, DM:PS, and dilution

system) ............................................. 14
4 Schematic of the Cohetester ................... ............. 16
5 Log-probability plot of coal particle-size distributionbased on sieve

analysis ............................................. 20
6 Respirable mass emissions and pressure drop as a function of time for Big
Brown coal at an AiC ratio of 4 ft/rain, Ryton fabric, and no conditioning.
Bags were cleaned every 60 minutes after an initial 180-minute period ...... 26
Brown coal at an A/C ratio of 4 ft/rain with Ry_n fabric and 50125 ppm of
1_3/503. Bags were cleaned every 60 minutes aP_r an initial 180-minute
period ............................................... 27
Brown coal at an A/C ratio of 4 R/rain with Ryton fabric and 24/12 ppm of
NH3/SO3. Bags were cleaned every 60 minutes after an initial 180-minute
period .............................................. 28
9 Respirable mass emissions and pressuredrop as a function of time for Big

Brown coal at an AJC ratio of 4 ft/min with Ryton fabric and 1216ppm of
NHJSO3. Bags were cleaned every 60 minutes _ an initial 180-minute
period.............................................. 29
I0 Pressuredropforbaseline andconditioning tests withBigBrown coalwith

Ryton fabric at an AiC ratio of 4 ft/min. Bags were cleaned every
60 minutes after an initial 180-minute period ...................... 30
Brown coal at an A/C ratio of 6 ft/min with Ryton fabric and no
conditioning. Bags were cleaned every 30 minutes during the last 4 hours
of the test ........................................... 31
LIST OF HGURES
Figure Page
Brown coal at an A/C ratio of 6 ft/rain with Ryton fabric and 24/12 ppm of
NHJSO3. Bags were cleaned every 30 minutes after an initial 90-minute
period .............................................. 32
13 Respirable mass emissions and pressure drop as a function of time for Eig
Brown coal at an A/C ratio of 8 ft/rain with Ryton fabric and 24/12 ppm of
NH3/SO3. Bags were cleaned every 30 minutes after an initial 90-minute
period ............................................... 33
14 Pressure drop for baseline and conditioning tests with Big Brown coal with
Ryton fabric at an AJC ratio of 6 ft/mim Bags were cleaned every
30 minutes during the last 4 hours of tests ....................... 34
15 Pressure drop for conditioning tests with Big Brown coal with Ryton fabric
at A/C ratios of 4 and 8 ft/rain. After initial fdtration periods, bags were
cleaned every 30 minutes at 8 ft/rain and every 60 minutes at 4 ft/rain ...... 34
16 Comparison of submicron emissions measured with the CNC for Big

Brown coal with Ryton fabric for the baseline and conditioning tests at an
A/C ratio of 6 f't/min .................................... 36
17 Respirable mass emissions and pressure drop for baseline and conditioning
tests with Big Brown coal with woven glass fabric at an A/C ratio of
4 ft/min. Bags were cleaned every 45 minutes after an initial 90-minute period 38
18 Respirable mass emissions and pressure drop as a function of time for

Milner coal at an AJC ratio of 4 ft/rain with Ryton fabric and no
conditioning. After an initial 90-minute filtration period, bags were
cleaned every 30 minutes, except that the last two cleaning cycle intervals
were 60 minutes ....................................... 39

Milner coal at an AJC ratio of 4 ft/min with Ryton fabric and 24/12 ppm
of NH3/SO3. After an initial 90-minute f'tltration period, bags were cleaned
every 30 minutes, except that the last two cleaning cycle intervals were
60 minutes ........................................... 40
20 Respirable mass emissions and pressure drop as a function of time for Milner
coal at an A/C ratio of 4 ft/min with Ryton fabric and 12/6 ppm of NH3/SO3.
After an initial 90-minute filtration period, bags were cleaned every 30 minutes,
except that the last two cleaning cycle intervals were 60 minutes .......... 41
iii
"°
LIST OF FIGURES
21 Respirable mass emissions and pressure drop as a function of dme for

M.finercoal at an A/C ratio of 8 ft/min with Ryton fabric arid24/12 ppm
of NH3/SO3. After an initial 40-minute filtrationperiod, bags were cleaned
every 20 minutes ....................................... 43 t

Milner coal at an A/C ratio of 4 ft/rain with P84/Ryton fabric and no
conditioning. After an initial 90-minute filtration period, bags were
cleaned every 30 minutes .................................. 44
23 Respirable mass emissions and pressure drop as a functionof time for

Milner coal at an A/C ratio of 4 ft/rain with P84/Ryton fabric and
24/12 ppmof NH3/SO3. After an initial90-minutef'fltraEonperiod,bagswere
cleanedevery30 minutes .................................. 45
24 Particle-sizedistributionof Big BrownandMilner fly ashbasedon

multicycloneresults ..................................... 47
25 Tensilestrengthasa functionof porosityfor baselineandconditionedfly

ashwith Big Brownand Milner coalswith Rytonfabricat an A/C ratio of
4 ft/rain ............................................. 47
26 Tensile strengthas a function of porosity for Task 2 Big Brown tests ....... 48
27 Tensile strength as a function of porosity for Task 2 Milner tests ......... 48
28 Effect of conditioning on baghouse pressure drop for 100-hour Big Brown

tests for Runs 403, 404, and 405. Bags were pulsed once per hour ........ 53
29 Effect of conditioning on average respirablemass emissions for 100-hour

Big Brown tests for Runs 403, 404, and 405 ...................... 55
30 Respirable mass emissions before and after bag cleaning for the 100-hour
baseline Run 403 with Big Brown coal .......................... 55
conditioning Run 404 with Big Brown coal ....................... 56
conditioning Run 405 with Big Brown coal ....................... 56
33 Range of submicron particle emissions for the 100-hour baseline Run 403
with Big Brown coal ..................................... 57
iv
LIST OF FIG_
34 Range of submicron particle emissions for the 100-hour conditioning

• Run 404 with Big Brown coal ............................... 57

Run 405 with Big Brown coal ............................... 58
36 Effect of conditioning on baghouse pressure drop for 100-hour Milner tests

for Runs 406, 407, and 411. Bags were pulsed every 30 minutes ......... 60
37 Effect of conditioning on average respirable mass emissions for 100-hour

Milner tests for Runs 406, 407, and 411 ......................... 61
baseline Run 406 with Milner coal ............................ 61
conditioning Run 407 with Milner coal ......................... 62
40 Respirable mass emissions before and after bag cleaning for the lO0-hour
conditioning Run 411 with Milner coal ......................... 62
41 Range of submicronparticle emissions for the 100--hourbaseline Run 406

with Milner coal ....................................... 63

Run 407 with Milner coal .................................. 64
43 Range of submicron particle emissions for the 100-no,,Tconditioning

Run 411 with Milner coal .................................. 64
44 Effect of conditioning on baghouse pressure drop for 100-hour Black Thunder

tests for Runs 409, 410B, and 413. Bags were pulsed once every hour ...... 66
45 Effect of conditioning on baghouse pressure drop for Black Thunder test

with 24/12 ppm of NH3/SO3for Run 410A. Bags were pulsed once per
hour for the f'LrSt 8 hours and once every 30 minutes for the remainderof
the test ............................................. 67
46 Effect of conditioning on average respirable mass emissions for 100-hour

. Black Thunder tests for Runs 409,410B, and 413 ................... 68
baseline Run 409 with Black Thunder coal ....................... 68
LIST OF FIGURF._
Fibre
conditioning Run 41OB with Black Thunder coal .................... 69
conditioning Run 413 with Black Thunder coal ..................... 69
50 Range of submicron particle emissions for the 100-hour baseline Run 409
with Black Thunder coal .................................. 70
91 Range of submicron particle emissions t'or the 100-hour conditioning

Run 410B with Black Thunder coal ............................ 70

Run 413 with Black Thunder coal ............................. 71
• 53 Respirable mass emissions for Black Thunder test with 24/12 ppm of
NH3/SO3, Run 410A ..................................... 71
54 Range of submicron particle emissions for Black Thunder test with

24/12 ppm of NH3/SO3, Run 410A ............................ 72
55 Effect of conditioning on baghouse pressure drop for 8-hour Big Brown

tests with a fun dust loading at an A/C ratio of 16 ft/rain for Runs 399 and
400. Bags were pulsed every 5 minutes during the last 2 hours of each test ... 73
56 Effect of conditioning on average respirable mass emissions for 8-hour Big

Brown tests with a full dust loading at an A/C ratio of 16 ft/min for
Runs 399 and 400 ...................................... .75
baseline Big Brown test with a full dust loading at an A/C ratio of
16 ft/rain for Run 399 .................................... 75
58 Respirable mass emissions before and after bag cleaning for the 8-hour Big
Brown test with conditioning and with a full dust loading at an A/C ratio of
16 ft/min for Run 400 .................................... 76
baseline Big Brown test with a reduced dust loading at an A/C ratio of
16 ft/rain for Run 401 .................................... 76
vi
LIST OF FIGURES
60 Respirable mass emissions before and after bag cleaning for the 8-hour Big
Brown test with conditioning and a reduced dust loading at an A/C ratio of
16 ft/rain for Run 402 .................................... 77
61 Particle-size distribution of Big Brown fly ash based on multicyclone sampling . 82
62 Particle-size distribution of Milner fly ash based on multicyclone sampling .... 82
63 Comparison of particle-size distributions of Big Brown, Milner, and Black

Thunder fly ash based on multicyclone sampling .................... 83
64 Comparison of particle-size distributions for normal and reduced dust-

loading tests for Big Brown, based on multicyclone sampling ............ 83
65 Effect of conditioning on fly ash tensile strength for 100-hour Big Brown
tests ............................................... 86
66 Effect of conditioning on fly ash tensile strength for 100-hour Milner tests .... 86
67 Effect of conditioning on fly ash tensile strength for 100-hour Black

Thunder tests ......................................... 87
68 Comparison of fly ash tensile strengthsfor conditioned tests with Big

Brown, Milner, and Black Thunder coals ........................ 87
69 Comparison of fly ash tensile strengths for baseline tests with Big Brown,
Milner, and Black Thunder coals ............................. 89
70 Fly ash tensile strengths for 100- and 8-hour Big Brown tests ............ 89
71 Fly ash tensile strengths for 100- and 8-hour Milner tests .............. 90
72 Comparison of tensile strengths of hopper ash and residual dust cake ash
from the bags for Black Thunder tests .......................... 90
73 Laboratory measurement of K: as a function of porosity with Carman-

Kozeny equation fit to data for Big Brown tests .................... 92
74 Laboratory measurement of K, as a function of porosity with Carman-

Kozeny equation fit to data for Milner tests ....................... 92
75 Laboratory measurement of K: as a function of porosity with Carman-

Kozeny equation fit to data for Black Thunder tests .................. 93
vii
LIST OF FIGURES
°
F_mn
76 Comparison of K2curves for Big Brown, Milner, and Black Thundertests .... 93
77 Correlation between initial K2 and aerated or packed porosity 95
78 Correlationbetween effective K2and aerated or packed porosity .......... 95
79 Residual dust cake at the end of the 100-hourbaseline test with Big Brown
coal for Run 403 ...................................... 100
80 Residual dust cake at the end of the 100-hourconditioning test with Big
Brown coal at an A/C ratio of 4 ft/rain for Run 404 ................ 101
81 Residual dust cake at the end of the 100-hourconditioning test with Big
Brown coal at an A/C ratio of 6 f't/minfor Run 405 ................ 102
82 Residual dustcake at the end of the 100-hourbaseline test with Milner coal
for Run 406 ......................................... 103
83 Residual dustcake at the end of the 100-hourconditioning test with Milner

coal at an A/C ratio of 4 ft/rain for Run 407 ..................... 104
84 Residual dust cake at the end of the 100-hour conditioning test with Milner
coal at an A/C ratio of 6 ft/rain for Run 411 ...................... 105
85 Residual dust cake at the end of the 100-hourbaseline test with Black
Thunder coal for Run 409 ................................ 106
86 Residual dust cake at the end of the 100-hour conditioning test with Black
Thunder coal at an A/C ratio of 4 ft/min for Run 410B .............. 107
87 Residual dust cake at the end of the 100-hour conditioning test with Black
Thunder coal at an A/C ratio of 6 ft/min for Run 413 ............... 108
viii
LIST OF TABLES
I Matrix for Task 2 ...................................... 7
2 Matrix for Task 3 ...................................... 8
3 Pulse-Jet Fabric Filter Specifications ........................... I1
4 ACGIH Respirable Mass Del'tuition ............................ 13
5 Summary of Coal Analyses ................................. 19
5 Results of Task 2 Testing .................................. 23
7 Physical and Chemical Fly Ash Characteristics from Task 2 ............. 24
8 Results from Task 3 Testing ................................ 51

i
9 Physical and Chemical Fly Ash Characteristics from Task 3 ............. 52
10 Ammonia and Sulfate Balances from Task 3 ...................... 96
ix
PULSE-JET BAGHOUSE PERFORMANCE IM]_OVEMENT WITH FLUE GAS
CONDITIONING: FINAL PROJECT REPORT
• 1.0 PROJECT SUMMARY
This project was jointly funded by the U.S. Departmentof Energy (DOE), the Electric
Power Research Institute (EPRI), and the CanadianElectrical Association (CEA), with
approximately50% of the fundingfrom DOE and 25 % each from EPRI and CEA. The
purpose of this project was to evaluate the effectiveness of flue gas conditioning in improving
the performance of pulse-jet baghouses. The work was conductedat the Energy and
Environmental Research Center (EEOC) at the University of North Dakota.
The project was divided into three tasks. Task 1 included facility preparation,
procurement of supplies, project management,and report preparation. Task 2 consisted of
15 screening tests of about 8 hours each for initial evaluation of the test pa:ameters. Task 3
• consisted of five 8-hour tests at an A/C ratio of 16 ft/rain and 9 longer-termtests of about
100 hours each to establish the effect of conditioning over multiple cleaning cycles. Task 2
results were presented in a Task 2 SummaryReportin September1991, and Task 3 _sults
were presented in a Task 3 SummaryReport in May 1992. This final report includesthe
results from both Tasks 2 and 3.
Previous research with shaker-cleanedand reverse-airbaghouses has shown that

particulate emissions and baghousepressure drop are significantly reduced when small
amounts of ammonia and SO3are injected into the flue gas just upstreamof the baghouse
(I-10). This concept also appearsto have significantpotential to improve the performance of
pulse-jet baghouses. The benefits of using flue gas conditioning with pulse-jet baghouses
may be even more significant than with reverse-gas baghouses because puLse-jetbaghouses
typically operate at higher air-to-cloth (A/C) ratios, are more subject to high particulate
emissions, andaremoresubject toa combination ofhighpressure drop,frequent bag
cleaning,
and short
baglife.
Thereareatleast
fouradvantages
ofconditioning,
allofwhich
aredesirable
forpulse-jet
baghouses:
" I. Particulate
emissions
aresubstantially
reduced,
whichenhances
theability
ofthe
baghouse
tomeetregulatory
standards
athighA/C ratios.
2. A more porous dust cake is formed, leading to reducedpressure drop. This means
that the baghouse can be operatedat reducedpressure drop while keeping the same
cleaning cyc!e, or the pressure drop can be held constantat a reducedbag-cleaning
frequency, or the baghouse can be operatedat a higher A/C ratio without
increasing either pressure drop or bag-cleaning frequency.
3. The dust cake is kept on the surface of the fabric. Fabric blinding by seeping and
packing into the fabric structureis prevented.
4. Dispersion of the ash afterpulsing is minimized because the dust cake is more
cohesive. The dust is dislodged from the bag in larger agglomerateswhich easily
fall to the hopper instead of being recollected on the bags.
The work was completed with KERC facilities consisting of a pulverizedco_-fired

combustor and pilot baghouse. Full-scale pulse-jet bags were employed underconditions
similar to large-scale baghouses. Tests included baseline runs to establish performance
without conditioning and similar mns with conditioning agents to quantifyperformance
improvement. The primary independentvariables included coal type, conditioning agent
concentrations, A/C ratio, and fabric type. The main dependentvariables were particulate
emissions, baghouse pressure drop, and cohesive p_ of the fly ash.
Results from the fifteen 8-hour tests completed in Task 2 showed that:
• Conditioning significantly reducedpressure drop and corresponding I_ at ali

concentrationstested. Concentrationsof 24/12 ppm of NHs/SO3provided
significant performance improvementfor the coals, fabrics, and A/C ratios tested.
• Conditioning did not reduce pani_mlateemissions for either the Ryton or

P84/Ryton fabrics at an A/C ratio of 4 ft/rain with a constant bag-cleaninginterval.
However, collection efficiency was already high at 99.99%, and some additional
reduction in particulateemissions would be likely if the bags were cleaned on a
pressure drop basis. The 8-hour tests were apparentJynot long enough in duration
to detect differences in particulate emissions with the felt fabric, however, the 100-
hour tests resulted in significantly reduced emissions with conditioning. A small
reduction in particulate emissions was seen with conditioning for the Ryton fabric
operated at an A/C ratio of 6 ft/min. Conditioning resulted in a significant
reduction in particulate emissions for the woven glass fabric.
* With conditioning, reasonablepressuredrop and excellent particulatecollection
efficiency were achieved for both the Big Brown and Miiner coals when operating
the pulse-jet baghouse with Ryton fabric at a high MC ratio of 8 ft/mim Results
imply that operation at even higher A/C ratios shouldbe possible when using
• ammonia and SC_ conditioning.
• The primary effect of conditioningon ash properties appearsto be a significant

increasein aeratedandpacked porosity. Conditioningalsoproduces an increasein
the tensile strength at constantporosity. It appearsthat reduced particulate
emissions axtdpressure drop are a direct result of the change in the cohesive nature
of the ash that occurs with conditioning.
Results from the 100-hour tests at A/C ratios of 4 and 6 ft/rain and the 8-hour tests at
an A/C ratio of 16 ft/rain completed in Task 3 showed that:
• Conditioning significantly reduced pressure drop and th_ correspondingspecific

dust resistance coefficient, K2, for the Big Brown and Milner coals at
• concentrationsof 24/12 ppm of NI-I3/SO_.With r_constant bag-cleaningfrequency,
the baghouse ziP was lower with conditioning at an A/C ratio of 6 ft/min than for
the correspondingbaseline tests at 4 ft/mim
• Conditioning reducedparticulateemissions for the Big Brown and Milner coalsat

concentrationsof 24/12 ppm of Nt_SO_ using Rymn or P84/Ryton fabrics at A/C
ratios of 4 and 6 ft/min with a constantbag-cleaning interval.
• The effects of conditioning on ash properties for the Big Brown and Miiner coals
were significant shifts in the tensile strength-porositycurves and increases in the
aerated and packed porosities. There were no obvious problems with bag
cleanability.
• Conditioning reduced pressure drop and correspondingK: for the Black Thunder
coal at concentrations of 12/6 ppm of NH3/SO3. Particulate emissions were not
reduced, but emissions were already very low for the baseline test.
• The ammonia slip was typically 1 ppm or less for the Big Brown and Milner tests,
but was 5 ppm for the Black Thundertests. The higher ammonia slip and smaller
changes in the tensile strength and porosity of the fly ash indicate that achieving
optimumperformance improvement with conditioning may be somewhat more
difficult with the Black Thunder coal.
• For Big Brown coal with a normaldust loading and off-line cleaning at an A/C
ratio of 16 ft/min, conditioning will lower the Ks. However, a more energetic
pulse may be required to sufficiently dislodge the dust.
• BigBrowntestswitha dustloading
ofapproximately
1/20thofnormalat16ft/min
withoff-line
pulsing
resultedinbagblinding.
Neitherconditioning
northeuseof
Neutralite*
precoating
was successful
inpreventing
thebagblinding. The
combination
ofverylow dustloading
andhighvelocitymay allowthedusttopack
tightly
intothefabric
structure,
ortheparticulate
properties
may changebecauseof
different
combustion
conditions.
These results demonstrate the significant benefits of using conditioning with a pulse-jet
baghouse: a substantialreduction in particulateemissions and a substantialreduction in
pressure drop (or the ability to operate at a higher A/C ratio without in_g pressure drop
or bag-cleaning frequency). The improvementsin fabric filter performance correlate strongly
with a shift in the tensile strength-porositycurves and with increases in the aerated and
, pac_ced porosity oftheflyash.
Conditioning appearsto be applicable to a wide range of coals and fabrics.

Applications for this technology are where t.tm_ is a need to reduce pressuredrop and/or
particulate emissions in existing baghouses, to reduce f'me-pa_cle air toxic emissions which
may be required in the future, and for new baghonse installationsto allow operation at a
higher A/C ratio while providing an ultrahigh fine-particle collection efficiency.
2.0 PROJECT OBJECTIVF_
The primary objective of the work was to evaluate the effectiveness of flue gas
conditioning in reducing mbe sheet pressure drop and particulate emissions from a pulse-jet
fabric f'dter. Specific goals of the work included the following:
1. Evaluate the effectiveness of conditioning _¢ithpulse-jet cleaning for three different

coals
2. Determine the effect of conditioning agent concentrationson baghousepressure
drop and particulateemissions
3. Evaluate the potentialto increase air-to-cloth ratio with flue gas conditioning
4. Determine the effectiveness of conditioning for two different fabrics
5. Demonstratethe effectiveness of conditioning over multiplecleaning cycles in 100-

hour tests
3.0 SCOPE OF WORK
The wvrk was inCmdedto be sufficient in scope to provide data that could serve as a
basis to determine the advantagesof using flue gas conditioning with full-scale, pulse-jet
baghouses. The primary independentvariables in the test program, in order of importance,
were:
1. Coal Type: While conditioning was pre,_iouslyshown to be effective for several

coals, the level of performanceimprovementwill likely be coal-specific. The first
test coal was Big Brown Texas lignite. The second coal was from the Alberta
Power Milner Station. The third coal was from the Black Thundermine in
Campbell County, Wyoming.
2. Conditioning Agent Concentration: Previous tests at the EERC have shown that,
for several coals, 25-ppm ammoniaand 12-ppmSO3are adequateto provide
substantialreductionsin pressure drop and particulateemissions. These
concentrations, however, have not been optimized for any coal and are likely to be
different for different coals. Therefore, three initialNH3/SO3concentrations of
50 ppm/25 ppm, 24 ppm/12 ppm, and 12 ppm/6 ppm were tested in Task 2.
Concentrationsof conditioning agents were not a primaryvariablein Task 3, but
were tested at two levels for the Black Thundercoal.
• 3. Air-to-Cloth Ratio: Since the economics of baghouses are highly dependent on

A/C ratio, a goal of the work was to determine ff A/C can be substantially
increased with conditioning. A/C ratios of 4, 6, and 8 ft/rain were tested in
Task 2. The 100-hour tests included A/C ratios of 4 and 6 ft/min, and additional
8-hour runs included tests at 16 ft/min.
4. Fabric Type: Ryton relied fabric (18 oz/yd2 with Ryton scrim) was the main test
fabric. Some 8-hour tests were also conducted with 22-oz/ya¢ woven glass (3PS
Style ?577 with acid-resistant finish). For the 8-hourMilner coal tests, P84FRyton
felted fabric was used as the second test fabric in place of the woven glass fabric.
Ali of the 100-hour and additional8-hour tests in Task 3 were conducted with
either Ryton or P84FRytonneedled felt material. New bags were used for each
test.
The original plan was to conduct fifteen 8-hour tests in Task 2 and nine 100-hour tests
in Task 3. Following the completion of the Task 2 tests, a review meeting was held at the
EERC to discuss initial results and plan any modifications to the originallyproposed scope of
work. One of the requestedchanges discussed in the review meeting was to conduct
additional8-hour _ at an A/C ratio of 16-ft/rainwith Big Brown coal. To achieve an A/C
ratio of 16 ft/min, only one 10-ft Ryton bag was used in the pilot baghouse. The 16-ft/rain
tests included baseline (no conditioning) and conditioning tests at the normaldust loading and
with a reduced dust loading. One of the purposes of these tests was to simula_ conditions
where a high-ratiopulse-jet baghouse is _ downstreamof an KSP as in the EPRI
Compact Hybrid PL,ticulate Collector (COHPAC) concept. Tne normal d_t-l_dln_ te_
were conducted in a manner similar to the o0_-r tests, except the A/C ratio was 16 ft/min
and the bag was pulsed off-line instead of the normal on-line pulsing. To achieve a nxluced
dust loading, the combustor was fired with both mmml gas and coal with about 1/20rh of the
normal coal feed rate, while m_intRiningthe same fine gas.flow rate.
The test matrix for the 8-hour, Task 2 tests is shown in Table 1, and the test matrix for
the Task 3 tests is shown in Table 2. Originally, nine 100-hourtests were planned, with
four additional 8-hour tests at 16-ft/rain. The last 100-hour test was to have been a run with
woven glass fabric. However, because of remaining questions from the first four 16-ft/rain
tests, additional 16-ft/rain tests were attemptedin place of the run with woven glass. When
the last 16-ft/mi_ tests were unsuccessful, the remainder of the run was used to perform one
additional test with the Black Thunder coal using Ryton fabric.
TABLE 1
Matrix for Task 2 (8-hour tests)
• A/C
Test NH3/SO3 Ratio

No. Run Purpose Coal (ppm) (ft/rain) Fabric
1 379 Establish Initial Big Brown None 4 Ryton

Baseline
2 380 Evaluate Conditioning Big Brown 50/25 4 Ryton

Agent Concentration

Agent Concentration

Agent Concentration
Y 383 A/C Tests Big Brown None 6 Ryton
6 384 A/C Tests Big Brown 24/12 6 Rymn
7 385 A/C Tests Big Brown 24/12 8 Ryton
8 386 Fabric Tests Big Brown None 4 Woven

Glass
9 387 Fabric Tests Big Brown 24/12 4 Woven

Glass
I0 388 Establish Initial Milner None 4 Ryton

Baseline
11 389 Evaluate Conditioning Milner 24/12 4 Ryton

Agent Concentration
12 390 Evaluate Conditioning Milner 12/6 4 Ryton

Agent Concentration
13 391 A/C Tests Milner 24/12 8 P84/Ryton
14 392 Fabric Tests Milner None 4 P84/Ryton
15 393 Fabric Tests Milner 24/12 4 P84/Ryton

TABLE 2
Matrix for Task 3 (100-hour and 16-ft/raintests)
AIC
Test NI'I3/SO3 Ratio
No. Run Purpose Coal (ppm) (ft/rain) Fabric .
i t
16 399 Baseline, High A/C Big Brown None 16 Ryton
17 400 Conditioning, High Big Brown 24112 16 Ryton

AIC
18 401 Baseline, Low Dust- Big Brown None 16 Ryton

Loading
19 402 Conditioning, Low Big Brown 6/3 16 Ryton
Dust-Loading
20 403 Baseline Big Brown None 4 Ryton
21 404 Coal Effects Big Brown 24/12 4 Rymn
22 405 AIC Effects Big Brown 24/12 6 Ryton
23 406 Baseline Milner None 4 P84/Ryton
24 407 Coal Effects Milner 24112 4 P84/Ryton
25 409 Baseline Black None 4 Ryton

Thunder
26A 410A Coal and A/C Effects Black 24/12 6 Ryton

Thunder
26B 410B Coai Effects Black 12/6 4 Ryton

Thunder
27 411 A/C Effects Milner 24/12 6 P84/Ryton
28 412 High A/C, Neutralite® Big Brown 6/3 16 Ryton
29 413 A/C Effects Black 12/6 6 Ryton

, Thunder -
4.0 FACILITY DESCRIPTION AND PROCEDURE_
4.1 Particulate Test Combustor (Frc)
• The pilot furnace, known as the particulatetest combustor (Frc), is a 550,000-Btu/hr

pulverized coal-fired unit designed to generate fly ash representative of that produced in a
fuU-scale utilityboiler. The combustor isoriented verticallytominimize walldeposits. A
refractory lining helps toensure adequate flametemperature forcomplete combustion and
prevents
rapidquenching
ofthecoalescing
orcondensingflyash.The meanresidencetime
ofa particle
inthecombustorisapproximately
three
seconds,basedon thesuperficial
gas
velocity.
The coalnozzle
oftheFrc fires
axially
upwardfromthebottomofthe
combustor,
and secondary
airisintroduced
concentrically
totheprimary airwithturbulent
mixing.Inaddition, tertiaryairissupplied abovethebaseofthecombustor. Coalis
introduced to the primary air stream via a screw feeder and ejector. An electric air preheater
is used for precise Controlof the combustion air temperature. Water-jacketedheat
exchangers provide flue gas temperaturecontrol to the baghouse. The Frc instrumentation
' permits system temperatures, pressures, flow rates, flue gas constituentconcentrations,and
baghouse operating data to be monitoredcontinuously and recorded on a data logger.
Figure 1 illustrates the PTC schematically.
Hue gas samples were taken at three system samplepoints: the furnaceexit and the
baghouse inlet and outlet. After passing throughsample conditionersto remove the
moisture, the flue gas was analyzed for 02, CO2,SO2, NOx, and CO. Except for the CO:
and CO, each constituent was normally analyzed at both the furnaceexit and outlet of the
baghouse simultaneously, using two analyzers. The concentrationvalues from ali of the
: instrumentswererecorded continuouslyusingcircle charts.Inaddition, alldatawere
manually recordedatsettimeintervals. NOz was determined using twoThermoElectron
Chemiluminescent NOz analyzers. The O2,CO, and CO: analyzers aremade by Beckman,
" and the SO. analyzers are manufactured by DuPont. Each of these analyzers is regularly
calibrated a_d maintained to provide accurate flue gas concentration measurements.
i
The baghouse vessel is a 20-inch-ID chamber that is heat-traced and insulated. Flue
• gas is introduced near the bottom of the baghouse. Since the combustor produces about
: 200 acfm of flue gas at 300*F, three 13-ft by 5-inch bags were used for tests at an A/C ratio
of 4 fttmin. For tests at an A/C ratio of 6 ft/rain, one bag was removed to provide the
correct filtration area. For tests at an A/C ratio of 8 ft/rain, two 10-ft by 5-inch bags were
used, and one 10-ft by 5-inch bag was used for the 16-ft/rain tests. Each bag was cleaned
Pulse-Jet
Baghouse
Combustor
Elec.
Preheat _ Bypass
FD Fan Tert.Air
Sec. Air - 6
Prim. Air ! NH: SOs

' Injection Injection
Figure 1. Combustion system used for tests, showing the Frc and baghouse.
separately
withitsown diaphragmpulsevalve.The bagswere cleanedon timedintervals,
rather than cleaning initiated by pressure drop, in order to quantify differences in pressure
drop with different test conditions. Once bag cleaning was initiated, all three bags (or two at
A/C ratios of 6 or 8 ft/rain) were pulsed in rapid succession on=line. Baghouse specifications
are given in Table 3.
4.2 Particulate Sampling and Measurement
Near real-mc particulate measurements were conducted with an aerodynamic particle

sizer (APS 33) manufactured by TSI, Inc., shown schematically in Figure 2. The primary
10
TABLE 3
Pulse-Jet Fabric Filter Specifications
' Baghouse Temperature 300*F

Fabrics' 1. Ryton needled felt with Ryton scrim, singed face of
fabric ha and plain side out, weight - 17.56 oz/yd _,
permeability - 39.4 elm/fta @ 0.5" WG, Mullen
Burst - 496 psi
2. Woven glass 3 x 1 double-filling-face twill, weight-
22.36 oz/yd:, permeability - 27.1 cfm/f-d@ 0.5"
WG, Mullen Burst - 972 psi, loss on ignition -
7.1%,/PS style 7577 with "Acid Gard" finish
3. P84/Ryton felt, 15 oz/yd 2, from Albarrie
Environmental Services
Bag Size 5 inch by 13 ft @ A/C -- 4 and 6 ft/rain

5 inch by 10 ft @ A/C - 8 and 16 f't/min
Cages Royal Wire Products Inc., 20 vertical wires; 6-inch

ring spacing; 11-gauge pregalvanized wire
Bag-Cleaning Mode On-line @ 4, 6, and 8 ft/rain

Off-line @ 16 ft/min
Venturi None
Pulse-Reservoir Pressure 40 psig
Pulse-Nozzle Diameter 0.5 inch
Pulse Volume Per Bag 0.3 scf
Properties of Fabrics 1 and 2 were measured by Grubb Filtration Testing Services, Inc.
• advantages of this system are high resolution and short sampling time. In the APS, particle-
laden air is passed through a thin-walled orifice, and because of their higher inertia, the
particles lag behind the gas. The velocity lag is uniquely related to the aerodynamic
diameter of the particles. Therefore, the aerodynamic diameter of a particle can be
determined by measuring the particle velocity as it exits from the orifice. To measure the
11
Piir ticlll-t.llden Glls
. *..*_ ,o.l* _' : ....... _L_
,oo...oo...o. :::=-.,
l Phot Otllultlplillr
k"": 1
Flowmeter _ II
. _ I O.mp-Q ,1
L_.J
Muffler
-
Vscuum
"r---'-'- [ ChamOer i
Assembly Pump
Figure 2. Schematic of aerodynamic particulate sizer.
panicle velocity, the APS employs a laser which is split into two beams. The light scarred
by a particle passing through these beams is collected, and two pulses are emitted, separated
by the time taken for the particle to cross the distance between the two beams. From this
time interval, which is measun_ electronically, the aerodynamic diameter is calculated. For
most applications, the particle-size distribution for particles ranging in size from 0.5 to
30 _m can be obtained within 20 seconds, giv;_g near real-time measurements. The particle-
size distribution can be obtained on the basis of either a number or mass concentration.
However, rather than looking at emissions of several particle sizes, fine-particle emissions
can be combined by using a calculated value of respirable mass. The American Council of
Govemmcn_l and Industrial Hygienists' (ACGIH) definition of respirable mass, as given in
Table 4, was used. The ACGIH definition was extrapolated and interpolated to calculate the
percentage at the midpoint o_ each channel for that particle size. The respirable mass from all
of the channels was added to obtain the total respirable mass. This provides a convenient
and effective method of plotting fine-particle emissions as a function of time.
12
d'
TABLE 4
ACGIH RespirableMass Definition
' Aerodynamic
Diameter Respirable Mass Fraction
(+,m) (%)
2.0 90
2.5 75
3.5 5O
5.0 25
IO.O 0
To determine the concentrationof submicronparticles (0.01 to 1.0 _m), a condensation

nucleus counter (CNC) was used. The CNC measures particle numberconcentrationby
using the fight-scattering technique. Since the dry submicronparticles are too small to be
easily detected by fight scattering, the particles pass through an alcohol vapor which
condenses on the particle, forming a droplet. Each droplet is large enough to scattera
detectable amount of fight when it passes through a fight beam. The droplet size is ne_ly
independent of the size of the originalparticle over a wide rangeof particle sizes, so the L
fight scattered is a function of numberconcentrationonly, not of size distribution. This

instrumentwas used independentlyas a continuous real-timemonitor of the number of fine
particles present in the flue gas. An impactor, prior to the gas streamentering the CNC,
removed the particles larger than 1 _m, so the CNC measuredthe concentrationof
submicron pm-ficles.
For both the APS 33 and CNC particle analyzers, a dilution system was employed.
Dilution of the flue gas is necessary to prevent moisture condensation in the sensors and to
reduce particle concentration, which, if too high, may exceed the maximumoutput of the
- CNC and may cause coincidence error in the APS. A schematic of the particulate-sampling
system with diluter is shown in Figure 3. The APS pump draws the sample into the diluter
volume. A second pump is used to draw off a portion of the flue gas to recycle it back as
dilution gas after it has been cleaned and the moisture removed. As shown in the figure, both
the APS and CNC were operated simultaneously using this dilution systen The maximum
13
Flowmeter
Diluter
Therrnocou=;e
Filter
Brinks I,--.,=actor
PumD Conclenser
OMPS
TSl Diluter
Bypass tor [ APS 33 I

Pump Control
CNC
Figure 3. Schematic of the particulate-samplingsystem (APS 33, DMPS, and dilution

system).
dilution that can be obtained with this system is about 10 to 1; ff higher dilutions are ne_:ied,
the system is operated in conjunction with a model 3302 TSI diluter. With both systems
operating, dilutions up to 1000 to 1 can be obtained. Typically, a differentialmobility
particle sizer DMPS (shown in Figure 3) is also used in conjunctionwith the CNC to provide
the submicron particle-size distribution. A DMPS sample requiresapproximatelya 25-
minute period when the particulateemissions are fairly constant. Since the APS and CNC
data indicated changes in dust concentrationsduring the last 25 minutes of most cleaning
cycles, valid DMPS measurements were not possible. However, the CNC was operated on a
continuous basis during most of the APS samplingperiods. During a test, APS and CNC
readings were taken at least every 10 minutes and mo_ often during the fast few minutes
after bag cleaning.
In addition to the CNC and APS, a modified EPA-Method 5 was used to provide dust
loadings at the inlet and outlet of the baghouse to determine the overall particulate collection
efficiency. Two outlet dust loadings were typically completed for each 8-hour test, and 7
14
outlet dust loadings were typically completed for each 100-hour test. Each outlet dust-
loading was conducted over at least two complete cleaning cycles and, therefore, included the
spike in emissions just after pulsing the bags. One EPA-Method 5 inlet dust loading was
taken for each 8-hour test, and 4 inlet dust-loadingmeasurementswere completed for each
100-hour test. In addition, one inlet multicyclone samplewas takenfor each 8- and 100-hour
test. The multicyclone provided another inlet dust loading measurementas weft as a
determination of the particle-size distributionfor particles smaller than 10/zm.
4.3 Sulfur Trioxide and Ammonia Injection
Sulfur trioxide (SO3) was produced with a catalytic SO_ generatorin which air and SOs
axepassed through a vanadium-catalyst-f'flled
reactor that isheated to850°F, thereby
oxidizing the S02 to SOs. Conversion of SOs m S03, measuredwith 502 analyzers, was
determinedto be 70% :i: 10%. Therefore, the accuracy of the injected SO3concentration
was about 4-10%. Ammonia (from a tank of anhydrous ammonia) was injected upstreamof
the SO3and controlled using a mass flow controller calibrated fnr ammonia. Concentrations
of ammonia and SO3in the flue gas were based on the flue gas flow rate, measured
downstream of the baghouse with an Annubarflow element.
4.4 Coal and Fly Ash Analyses
For each coal, standardultimate/proximate/Btu/ASTMash analyses were completed.

In addition, major elemental analyses of the baghouse hopper ash were completed for each
test. Conditioning tests included a determinationof the amount of ammoniaadsorbed on the
fly ash and the amount of ammonia in the flue gas downstreamof the baghouse.
Physical
characterization
oftheflyashwas necessary
todocument changesasa result
oftheconditioning
processandtoprovideanunderstanding
ofthemechanisms ofthe
conditioning
process.Important
properties
included
thebulkflyashparticle-size
distribution,
measuredby a Coulter
counter;
particle
density,
measuredby a helium
pycnometer;
and specific
surface
area,
measuredby a BET monosorbinstrument.
A key
physical
measurementisthedetermination
ofthecohesiveproperties
oftheflyash,sincethis
appears
tobe oneofthemainpredictorsofashbehaviorinfabric
filtration.
Two
" instruments
manufacturedby HosokawaMicronInternational
wereusedtoquantifycohesive
properties.
A powdercharacteristics
tester
was usedtomeasuretheaerated
andpacked
" porosities
oftheflyash,anda Cohetester
was usedtomeasurethetensile
strength
ofthe
bulkashasa function
ofcompressive
force
orporosity.A schematic
oftheCohetester
is
15
shown in Figure 4. It consists of a horizontal split cell 5 cm in diameter, with one-half of
the cell movable and the other half fixed. The cell is suspended so that it can be pulled apart
with minimal force when no sample is in the cell, thus minimizing any error due to external
frictional forces. When the strain motor is named on, the powder bed is pulled and extended
in the same direction as the tensile force. The displacement (extension) of the powder bed
and tensile stress are plotted on an x-y recorder. The maximum value on the vertical axis is
the cohesive tensile strength of the powder in grams of force per square centimeter (g_cm 2)
at that compaction force.
The following procedure is used for making a tensile strength measurement using the
Cohetester. Because relative humidity may aff_ the cohesive properties of fly ash, the
samples are prepared by sifting them through a 60-mesh screen and storing them overnight in
a humidity-controlled chamber at 10% relative humidity. An amount of sample, based on
experience, is weighed, carefully poured into the cell, and compacted by guiding the packing
weight carefully inw the extension tube. Additional weights (up to a maximum total
compaction weight of approximately 5 kg) are placed on the first weight until the lip of the
packing weight rests on the cell extension tube. With the sample properly compacted, the
weights and cell extension tube are removed. After zeroing the recorder pen and releasing
the securing knob, the strain motor is turned on, and the fracture curve is recorded.
Figure 4. Schematic of the Cohetester.
16
Compactionweights ranging from 320 to 5000 grams were used to compress the
samples, which correspondsto compaction pressures of 16 to 255 gJcm 2. The upper value
of the compactionweight is limited by the structuralintegrity of the suspendedcell. The
lower limit is the minimum weight necessary to compact the ash so that it will breakalong a
plane when it is pulled, giving a valid tensile strengthmeasurement. The range in porosity
for a given fly ash sample is determined by the range in compactionforce, which was the
• same for each sample. For a particularash sample, at least five tensile strength
measurements over the full range of compactionforce are required to plot the tensile streagth
as a function of porosity. The percent porosity of the powder cake for each test was
determined from the particle density (measured by helium-air pycnomecry),the sample
weight, and the cell volume occupied by the compacted powder (supplied by the
manufacturer).
The aerated bulk density is obtained by sifting an ash samplethrough a vibrating 22-
mesh screen into a 100-cre3 cup, so that dust overflows the cup edge. The excess dust is
scraped off with a knife edge, and the weight of the known volume of dust is measuredto
determine the bulk density. The packed density is determined by adding an extension to the
cup and filling the extension with additional sifted ash. The cup with the extension is then
placed in a mechanism that raises the cup about 1/2 inch and lets the cup fall against a stop.
This is done once per second for a period of 3 minutes. The cup extension is then removed
and the excess dust scraped off as before. There is no external compaction force on the dust
layer. Compaction is caused by the naturalsettling that occurs as the dust is shocked. Since
panicle density is available, aerated and packed bulk densities are reportedas aerated and
packed porosities.
5.0 RESL_TS
5.1 Coal Description and Analyses
Three different coals were used for the test program: Big Brown Texas lignite, Milner
coal, and Black Thunder subbituminous coal. A short description of each of the coals is
given below:
Big Brown Texas Lignite: The Big Brown station, owned and operated by TU
Electric, is a minemouth facility located near Fairfield, Texas, in Freestone County.
The plant consists of two 575-MW pc-fired boilers. Big Brown lignite is pan of the
17
Wilcox Formation found in east Texas and is characterizedby relatively low heating
value and high moisture content. In addition, past experiences at the EERC and at the
Big Brown Station indicated that this coal produces a fly ash that is somewhatdifficult
to collect in a fabric f'flter.
Milner Coal: The H.R. Milner Station, owned and operated by AlbertaPower Ltd., is
located in Grand Cache, Alberta. The plant has a 150-MW pc-fired boiler that bums a
unique coal. The coal consists of rejects and dewatered tailings left over after washing
a high-quality metallurgicalcoal from the Mclntyre mine, located nearthe plant.
Although this coal is characterized by a very high ash content (>40%), the heating
value of the coal is still reasonable: > 8000 Btu/lb. The Milner stationis equipped
with a pulse-jet baghouse and has had a history of poor baghouse performancewhile
firing this coal.
Black Thunder Subbituminous Coal: The Black Thunder mine is located in the
Powder River Basin 12 miles east of Wright, Wyoming, in CampbellCounty. This
mine produces more coal than any other mine in the United States, approximately30
million tons per year. Cml from the Powder River Basin is subbituminousin rank, is
very low in sulfur ( < 1%), and is a prime candidatefor utifities consideringfuel
switching to meet SO2emission requirements.
A summary of the coal analyses is given in Table 5, and a plot of the sieve analyses is
shown in Figure 5. The samples were composite samples from the l(X)-hourba._-lineruns.
The coals were sampled from the coal feeder about once every 4 hours and then combined
into a single composite sample for the entire run. The moLqurein these samples is typically
much less than the run-of-mine coal moisture because some drying occurs duringpulverizing
and some additional drying may occur during samplingfrom the warm location near the
combustor. The Miiner ash content was very high at 42.2%, compared to 14.6% for the Big
Brown coal, and 6.7 % for the Black Thundercoal, on a moisture-free basis. Even though
the Milner coal had a much higher ash content, the heating value was similarto the Big
Brown. Ash analyses show that there were significant differences in silica and calcium
concentrations among the three different ashes. Silica concentration ranged from a low of
31% for the Black Thundercoal to a high of 59 % for the l_er coal, while calcium
concentrations ranged from a low of 4 % for the Miiner coal to a high of 23%for the Black
Thunder coal. The three coals were pulverized to a nominal 70% minus 200 mesh (74 tlm);
however, the Big Brown and Black Thundercoals were coarser at 52 % minus 200 mesh
compared to 70 % minus 200 mesh for the Miiner grind.
18
TABLE 5
Summaryof Coal Analyses
_ Black Thunder
. PROXIMATE (% m0istm'e-fre_
Ash 14.60 42.18 6.70
Volatile Matter 46.90 16.55 46.23
• 38.50 41.27 47.07
Fixed Carbon
Coal Moisture 23.60 1.90 4.99
_ULTIMATE (% moLs'ture-free)
Carbon 61.45 51.14 69.52
Hydrogen 4.82 2.75 4.63

Nitrogen 1.02 0.73 0.95
Sulfur 0.88 0.42 0.48
Oxygen (diff.) 17.20 2.76 17.70

Ash 14.60 42.18 6.70
HEATING VALUE (moisture-free)
Btu/lb 10,513 8,523 12,011
XRF COAL ASH ANALYSES

(% cone. as oxides)
SiO2 46.5 59.0 31.2
Al203 15.7 23.8 16.7
Fe_03 4.5 5.5 5.8
Ti02 1.9 1.2 1.1
P,.05 0.3 0.4 1.1
CaO 17.6 4.3 22.9
MgO 3.5 1.7 7.2

N_O 0.4 0.5 0.9
K20 0.4 2.0 0.5
SO3 9.2 1.6 12.6
S_I__,V'F.,
TEST (% retained__ 4.67
60 mesh
100 mesh 21.53 6.91 13.66
140 mesh 11.81 11.38 13.99
170 mesh 5.49 --
200 mesh 8.34 11.75 16.12
__ 6.54 --
• 230 mesh
270 mesh 10.97 7.42 11.91
325 mesh 4.30 4.90 1.23
" 37.56 51.10 38.63
Pan
% < 200 mesh 52.83 69.97 51.77
19
9
98
9S
_" 80
70
6O !
30 - A :gig-Brown!! !
20 ' o• UlaclcThunclei"
Milner
I0- ' t
$_ i • i : ' i
30 40 SO60708090 200 300 400 500

I00
Particle
Size(microns)
Figure 5. Log-probabilityplot of coal particle-size distributionbased on sieve analysis.
5.2 Theoretical Considerations
The size of fabric filters and bag-cleaning frequency are determinedby pressure drop.
Assuming viscous flow, pressure drop across a fabric filter is given as:
At) = KrV + K2W.V + K2CV_1/7000 [_. 1]

whore:
Ap = differential pressure across baghouse tube sheet (inches of water)

Kf = fabric/dust resistance coefficient (inches of water-rain/ft)
V = face velocity or air-to-cloth ratio (ft/min)
K2 = specific dust cake resistance coefficient (inches of water-ft-rain/lh)
WR = residual dust cake weight (lb/ft2)
C = dust loading (gtxins/acf)
t = filtration time between bag cleaning (min)
20
The first term in Equation1 accounts for the pressure drop across the fabric. For a
new fabric, pressure drop across the fabric alone is generally negligible, but, in cases where
the dust packs permanently into the interstitialspaces of the fabric, this may be a significant
term. The second term in Equation 1 accounts for the pressure drop contributionfrom the
• permanent residual dust cake that exists on the surface of the fabric. In many cases, after
long-term Ol_mtion, this is a significant term. The third term in Equation1 accounts for the
• pressure drop contributionfrom the dustaccumulated on the bags sine., the last bag cleaning.
Ali of the dust characteristicsthat affect pressure drop am combined intoa single term
K2, the specific dust cake resistance coefficient. The primary factors that affect K2 are the fly
ash particle-size distributionand the porosity of the dust cake. Prediction of K2in terms of
measurable dust parameters has been attempted, but accurate predictionof the actual
operating K2 for any given bagbou_ is difficult. One of the simpler models is the Carman-
Kozeny relationship, which has a strong theoretical basis and, assumingmonosized spherical
particles, takes the form:
K_= 36k _ (1 - 0 / (P PpDb [F.q.2]
where:
K2 = specific dust calm resistance coefficient (see/ft); note: K2 can be converted to

inches of water-ft-rain/lhby multiplying by a factor of 311.6
k = Carman-Kozeny constant (-5) (dimensionless)
= gas viscosity Ob-soc/ft2)
= porosity (dimensionless void volume fraction)
pp = particle density Ob/ria)
D = particle diameter (ft)
The Carman-Kozeny equation was fitted to laboratory measurements of K2 (presented

later in this report). Typical Ks values for pc-flred fly ash range from about 1 to 20 inches
of water-ft-min/lb, but may, in extreme cases, cover a wider range. K2 values can be
reported as an initial value and as the effective value. The initial value is based on the linear
part of the At) versus time curve from the first filtration cycle, starting with new bags.
During the first f'dtration cycle, the At) contribution from the first term is negligible, and
there is no residual dust cake, so Ks can be readily obtained from the third term in Equation
1. The largest source of uncertainty is the dust loading, which will typically fluctuate at
least + I0 %. An effective K: value is based on a At) that is an average of the difference
21
between the before- and after-cleaning _Ps. Effective K2 is higher than the initial K2
because the bags generally do not clean uniformly. Immediately after cleaning, the gas flow
will take the path of least resistance, so a higher velocity will occur in areas that were more
thoroughly cleaned. This nonuniformvelocity will graduallydiminish until there is a
complete cake over the entire surface again. The net result is that the AP increase between
bag cleanings will generally be greaterthan the initial increase in AP during the ftrst
filtration cycle, even though the same amount of dust is added in a given time interval.
The primary factors that determine collection efficiency in a fabric filter are dust
properties such as particle size and cohesive characteristics, fabric properties such as the pore
size and fiber diameter,bag-cleaning frequency, and face velocity. To achieve high
collection efficiency, the pores in the filter media must be effectively bridged (assuming they
are larger than the average particle size). However, pore-bridging ability may deteriorate for
a given fabric when the face velocity is increased. Recent work that the EERCconducted
for DOE, where pore bridging was studied as a function of pore size, velocity, and
conditioning, showed that pore-bridging ability was highly dependent on all three of these
factors (I I). Therefore, increasing the A/C ratio is likely to result in higher particulate
emissions unless a superior fabric is employed, or the dust is modified with conditioning
agents to improve pore bridging at a higher velocity. When considering the effect of
conditioning on dust characteristics,if a very high collection efficiency can be maintained
even after increasing the face velocity, it indicates that the ability of the dust to bridge pores
and resist reenwainmentis greatly improved.
5.3 Task 2 Results (8-hour tests)
A summary of the test variables and results from ali fifteen 8-hour tests is shown in
Tables 6 and 7.
5.3.1 Big Brown Tests
The purpose of the fast four tests was to evaluate the effect of conditioning agent
concentrations on particulate emissions and pressure drop with Ryton fabric at an A/C ratio
of 4 ft/rain. In addition to a baseline test, identical tests were conducted at conditioning
agent concentrations of 50/25 ppm of NH3/SO3, 24/12 ppm of NH3/SO3, and 12/6 ppm of
NH3/SO3. For these tests, the bags were cleaned once an hour following an initial 3-hour
filtration period• Particulate collection efficiency was high at 99.9+ % for ali of the 4-ft/rain
tests, indicating that the Ryton fabric provides excellent filtration capabilities. No measured
22
23
24
reduction in particulate emissions occurred with conditioningbased on the APS, condensation
nucleus counter, or EPA-5 results. However, if the bags had been cleaned on a pressure
drop basis rather than on a time basis, some reduction would likely have been noted. Plots
of respirable mass and pressuredrop as functions of dme for Tests 1 'through4 are shown in
. Figures 6 through 9. Pressure drop was reduced approximately50% with 24/12 ppm, of
NH3/SO3compared to the baseline test, as shown in Figure 10. At concentrationsof
• 50/25 ppm, pressure drop was very close to concentrationsof 24/12 pl)m, and at
concentrations of 12/6 ppm, pressure drop was just slighOy higher. Since particulate
emissions were not significantly different and pressure drop was just slightly different for the
three conditioning tests, the concentrationsthat provided the best performanceappeared
uncertain. To help assess the differences, K2values were calculated by three different
methods: one calculation was based on the first three hours of fdtration time; the second
calculation was based on the average AP increase between bag cleaning duringthe last 5
hours of the test; the third calculation was also based on the _ during the last 5 hours, but
the dust loading was based on the amount of ash collected in the baghouse hopper ratherthan
on measured inlet EPA-5 dust loadings. Results of these Ks measurementsare reported in
Table 6 as an average of the three calculations along with the standarddeviation. From the
K_value of 4.8 inches of water-ft-min/lbfor concentrations of 12/6 ppm of NHJSO_,
compared to K2 values of 3.6 and 3.5 for the other concentrations,it appearsthat the
effectiveness of conditioning is somewhat diminished at concentrations of 12/6 ppm of
NH3/SO3.Similarly, theaerated andpackedporosity data(showninTable6)indicate that
theeffectiveness ofconditioning isreduced atthelowerconcentrations of12/6ppm of
NHJSO3. Therefore, concentrations of24/12ppm ofNH3/SO3werechosen forsubsequent
testsathigher A/C ratios.
Tests 5 through 7 consisted of baseline and conditioning runs at an A/C ratio of 6

ft/min and a conditioning test at an A/C ratio of 8 ft/rain. Respirable mass and pressure
drop plots are shown in Figures 11 through 13. Particulate collection efficiency was slightly
reduced for the baseline test at 6 ft/rain compared to the baseline test at 4 ft/rain, but was
still reasonably high at 99.96 %. Collection efficiency was somewhat better, with
conditioning at 99.98% and 99.99% for the tests at A/C ratios of 6 and 8 ft/min,
respectively. However, the effect of conditioning on pressure drop was significant, as shown
e
in Figure 14. For the baseline test, after an initial 90-minute filtration period, the cleaning
cycle was set at 45 minutes. Pressure drop was not adequately controlled, so the interval
was reduced to 30 minutes, which was still not sufficient. The interval was then reduced to
' 20 minutes, along with a slight increase in pulse duration. After several cleaning cycles, this
brought the before-cleaning pressure drop to below 4 inches. Since the pressure drop had
25
102 G__
i
! i
. :
10'_. i ,, i • r-i_...........
.i_..
' "! i l i
g _ i , i
10"1 ..... "_......... V ............... ,
: i I
10"l 1._-....... i"..............
10.4 ,_
° o
.111_
JLU'S i i
i i Big Browfi Texas Lignite
ii ! Ryton Fabiric
=
m.
i '
!
E :
_ ,
, 0
0 100 200 300
.-.--"..,
,-.....
400 500
Tbne (r_)
Figure 6. Respirable mass emissions and pressure drop as a function of time for Big Brown
coal at an A/C ratio of 4 ft/min, Ryton fabric, and no conditioning. Bags were
cleanedevery 60 minutesafter an initial 180-minuteperiod.
26
:. 102 7
, _
N
10o
10"1 !!!!!'.
i.!!!i!
...
10.2 .........................
i........................................
;....................
_..................
.-_
10.3 ...................,..................... ;............... _...............
10.4
10"s . .
Big Brow_ Texas Lig_te
Ryton Fal_ric i
8 ...................._i
...........................
:A/_--4_mm'.........
'..............'.......................
_q
tm _i 50 ppm NH 3 & 25 ppm SO3
o. 6 ......
•........._-'" .......................
""'*"_ .........
"'-_-"--'_'- ......"......
" "-- ..........
m.
_° 2 .............. i i L "....................
0
0 100 200 300 400 S00
Time (rain)
• coal at an AJC ratio of 4 ft/min with Ryton fabric and 50/25 ppm of NI-I3/SO 3.
Bags were cleaned every 60 minutes after an initial 180-minute period.
27
-,,iu2 .... ,
•n,,l .......................... )p ....................,

i ...........
! .....i....._ ........O-.i
' i
100 .............. ............................. 4....,
i
10-2 iiii iii'

.................
....... )
................ '
i ' +
1114 i .............. _"...... ""%-"
ln'S
AU
.... , i ,
Big Brow¢.Texas Ligmte

Ryton Fabric i_.._ ..... _._._.._._,
8 ......._.._......._;._........_._......,
AIC = 4 ft/rain
tl
=,',, 24 ppm NH3 & 12 pp:mSO:3

Z:
O
t..
,- !
i "
2 _I. ,.
.
..,.._._--._.-
i
i
0
+
100 200 300 400
,.'t $00
T_e (rain)
coal at an A/C ratio of 4 ft/rain with Ryton fabric and 24/12 ppm of NI-I3/SOa.
28
l°2 + + i
• 101
• + 10°
lo"_ ..............
i..................
_..................................
i..........
iio+ i 10.3 .......................................

i
104
10-s
• Big Browfi Texas LigNte
Ryton Fabric
,..._o.ooo.ooo_ooo.o i °..o.....oo_o.o.o**..._..o.o_,
,-, 8 .......................
"- AIC -- 4 ft/rain
_ •
.._ 12 ppm _ & 5 ppm SO3
•" 6
0
L..
• i..............
i i...................
i.....
I_ 20 ....................
+" "-_.....
0 100 200 300 400 500
Time (min)
• Figure 9. Respirable mass emissions and pressure drop as a function of time for Big Brown
coal at an A/C ratio of 4 ft/mi_ with Ryton fabric and 12/6 ppm of NH3/SO3.
• Bags were cleaned every 60 minutes after an initial 180-minute period.
29
10 | ' Big Brown Texas Lignite ....
Ryton Fabric
AIC = 4 ft/rain
8
• 24 ppm NH3 & 12 ppm SO3

= 6
.. 4
0 .
0 I00 200 300 400 $00
Time (min)
Figure 10. Pressure drop for baseline and conditioningtests with Big Brown coal with
Ryton fabric at an A/C ratio of 4 f0min. Bags w_re cleaned everj_60 minutes
after an initial 180-minuteperiod.
decreased with successive 20-minute bag clcanings, the intervalwas increased again to 30
minutes for the last 4 hours of the test. However, pressure drop again increased with
successive cleaning intervals and reached a maximumof over 10 inches by the end of the
test. This indicated that pressure drop could not be adequately controlled with a 30-minute
cleaning interval for the baseline ash at an A/C of 6 f't/min(see Figure 14). With 24/12 ppm
of NH3/SO3 and the same 30-minute cleaning cycle, pressure drop was easily controlled at
2.5 inches WC (see Figure 14), which shows that conditioning greatly facilitates operation at
high A/C ratios. Since difficulties were encountered for the baseline case at an A/C of 6
ft/rain, no baseline test was attempted at 8 ft/rain; however, a test with conditioning was
conducted at 8 ft/rain. Figure 15 is a comparison of pressure drop for the conditioning tests
with 24/12 ppm of NHJSO3 at A/C ratios of 4 and 8 ft/rain. At 4 f't/min,pressure drop was
about 2 inches with a 60-minute cle.xninginterval. Raising the A/C ratio from 4 to 8 ft/min
30
Time (min)
• Figure 11. Respirable mass emissions and pressure drop as a function of time for Big Brown
coal at an AiC ,:atio of 6 ft/rain with Ryton fabric and no conditioning. Bags
. were cleaned every 30 minutes during the last 4 hours of the test.
31
1o21
101 .......................
!'....i_
3.....,: i
r-"' :'-:
"- 7- "'
i '
"_ 10.3
1.o
"4 i....................
: i
: i
I0 "$ , , ..... : ....... _ , .....
iBig Brow_ Texas Lignite
A/C = 6 ft/rain !
24 ppm N __ & !2 Ptlm SO=
3
: .
0
0 100 200 300 400 500
Time (min)
coal at an #dC ratio of 6 ft/mm with Ryton fabric and 24/12 ppm of NH3/SO 3.
32
. 103
10_
._.._.,,..;...._._._..._..
II£ ; _I
b
I0 ° I.................... [_.... '_--i" "....
-_ - ®i,_ ''....... .... ....J'"-q
_
10 "1 "'-t_'-I ..... I ...._.................................
D
i
1o.2 ,|.7_.-J ' ,_.....
104 :::'"._L .
lO-S ....®....®.... _ --
104
Big Brown Texas Lignite '_
_i [ --AIC = $ f_Imin
mlll',l I
ml
II.
! _i...... 1_.. ,..... i I
o 4 .............. •.........
i , -.
! + 1
...........
0
0 100 200 300 400 500
Time (min)
, coal at an A/C ratio of 8 ft/min with Ryton fabric and 24/12 ppm of NI_3/SO 3.
33
• 24 ppm NI_ & 12 ppm $O 3
O Baseline I
.. 12 .....
Big Brown Texas Lignite
I0 .? Ryton
!|.... Fabric .................
A7C"-'_T_BiFn 'ICP" -
i
-,.. i ,o o.oooooo_o.oo.ooo-oooo...o-oo*_.ooo.o*-.o_ooo-o ooo' ) "*
6 .....'i ...........................................
)
0 i
0 100 200 300 400 500
Time (rain)
Figure 14. Pressuredropfor baselineandconditioningtestswith Big Browncoalwith
Ryton fabric at an MC ratio of 6 ft/min. Bags were cleaned every 30 minutes
during the last 4 hours of tests.
||:
8 -[ ....... ii Big BrownTexasLignite

• k/C = 4 f_mio RytonFabric
O A/C= 8 _mia 24 ppmNI'I3 & 12ppmSO3
6 1
_a g)
"'_
=
2 Z ID O)
_'"_"p........_........
'-
•
: 0 ' F,It+ l i
3+'
0 _ ' l ' I ' ' I
0 I00 200 300 400 500
Time (rain)
Figure 15. Pressure drop for conditioningtests with Big Brown coal with Ryton fabric at
A/C ratios of 4 and 8 ft/rain. After initial filtrationperiods, bags were cleaned
every 30 minutes at 8 f-t/rainand every 60 minutes at 4 ft/min.
34
should result in a pressure drop increase to 8 inches WC for the same cleaning interval and
K:. If the cleaning cycle interval is reduced by 50% at the same time, the pressure drop
should increase only to 4 inches WC. Figure 15 indicates that, when the A/C ratio was
doubled and the cleaning interval reduced by 50%, pressure drop increased from about 2 to 4
inches WC, as predicted by theory. This implies that K: did not increasewith a doubling of
velocity and that the bags cleaned weil, even at an A/C ratio of 8 ft/min. Actual K2
• measurements show that K2did not increase, but was in the range of 3.0 to 3.5 2:0.9 inches
of water-ft-rain/lb. Figure 15 does show a slight increase in _t' with time, indicating some
increase in residual dust cake weight. The extent of this trendwas furtherevaluated during
the 100-hour tests. Figure 13 includes a plot of respirable mass emissions for the test at an
A/C ratio of 8 ft/min. The large range in emissions from just after bag cleaning to just
before bag cleaning covers five orders of magnitudefrom about 10 mg/m3to 104 mg/m3.
Since the inlet respirable mass was about 600 mg/m3, this correspondsto a respirablemass
collection effficiencyfrom 98% just after bag cleaning to 99.99998% just before bag
cleaning. During two of the cleaning intervals, a continuous30-minutesample was taken, as
shown in Figure 13. Clearly, most of the emissions occur immediatelyfollowing bag
cleaning; however, two encouraging results are evident. The respirable mass spikes are
lower with time, mdicatingthat some residual dust cake is developing. This helps to
maintain pore bridges and results in a decrease in average emissions with time, indicating
that an even higher colle_ion efficiency can be achieved with longer-termoperation. During
the second 30-minute APS sample, the respirablemass collection efficiency agreed closely
with the EPA-5 collection efficiency, both at 99.99%. These results are highly encouraging
because they show that, when conditioningis employed with this fabric and coal, a pulse-jet
baghouse can be operated at a high A/C ratio of 8 ft/rain with an easily controlled pressure
drop and a very high collection efficiency.
Submicron emissions for ali of the tests closely followed the APS results. An example
of results of submicron emissions for the Big Brown tests with Ryton fabric at an A/C ratio
of 6 ft/min is shown in Figure 16. One difference between the APS and CNC results is the
peak emissions just after bag cle,aning. The submicronemissions with conditioning are
slightly higher than for the baseline case. An explanation is that the actual submicron
particle count at the inlet to the baghouse is increased with conditioning due to the formation
of f'me ammonium sulfate particles. Just after bag cleaning, when there is a minimum dust
¢
cake, collection efficiency for these particles is low. Therefore, a higher inlet concentration
results in a higher level of submicron emissions just after bag cleaning. However, after a
short time, these submicron emissions quickly drop to very low levels and are of no concern.
35
Ryton Fabric 0 r_aseline ]
107
A/C - 6 ft/min
....
• 24 ppm NI-I3 & 12 ppm SO3
f
_ l0 s
104 ......... 1....
.-_ 106
103 " _ . i' _ .... ._ ,
"_ 102
101
•_ l0 t
10"l
c_
10.2
106
." I0s
_ lOl
xo°
•_ 10"l
o_ 10"2
10"3 ------.e-----
0 100 200 300 400 500
Time (min)
Figure 16. Comparison of submicron emissions measured with the CNC for Big Brown coal
with Ryton fabric for the baseline and conditioning tests at an A/C ratio of
6 ft/min.
36
Pressure drop and respirable mass results with 22-oz/yd2woven glass fabric at an A/C
ratioof4 ft/min arcshowninFigure17. The effect ofconditioning withwovenglass fabric
isevenmorepronounced thanwithRytonfelted fabric• Pressuredropwas reduced from
about6 inches WC forthebaseline test
toabout2 inches WC withconditioning forthesame
" cleaninginterval. The outletEPA-5 dustloading was reduced by a factoroftenwith
conditioning. The peakrespirable massemissions were also abouttentimeslowerwith
" conditioning,andtheminimumemissions wereaboutthree orders
ofmagnitude lowerwith
conditioning. Forthebaseline test,
minimumemissions werehighataboutI mg/m3,which
isan indicationthatcomplete porebridging didnotoccurandthat pinholeswerelikely
present. The much lowerminimumemissions withconditioning indicatethatconditioning
providessuperior porebridging. Although no tests wereconducted withthewovenglass
fabricathigher A/C ratios, these resultsindicatethatincreasing theA/C ratio should be
possiblewhen conditioning isused.Forthebaseline case,on theother hand,remits indicate
thatpoorcollection efficiency wouldbe a problemfftheA/C ratio wereincreased forthis
coalandfabric. A collection efficiencyofonly99.6%wasachieved atan A/C ratio of
• 4 f't/min
withthebaseline ash,whichwouldnotmeettheNSPS particulate emission limitof
0.03Ib/106 Btu.Whilecollection efficiencymay improve withlonger-term operation, these
tests
flagthecombination ofwovenglass fabricwiththeBigBrown coalandpulse-jet
cleaningasa potential problem, butindicate theproblemcanbeeliminated with
conditioning.
5.3.2Miln¢rTests
A total of 6 tests were completed with the Milner coal. Compared to the Big Brown
tests, notable differences are the much higher inlet dust loading and the higher loss on
ignition ('LOD of the fly ash for the Milner coal (shown in Tables 6 and 7). The fly ash LOI
for the Big Brown coal was less than 1%for ali but one test, while the LOI for the Milner
ash ranged from 3.8 % to 6.6 %. This LOI range is typical of the LOI of the baghouse ash
that has been observed at the Milner station.
Three initial tests were completed at an A/C ratio of 4 ft/min with Ryton fabric, a
baseLinetest, and conditioning tests at concentrations of 24/12 ppm and 12/6 ppm of
• NH3/SO3. Respirable mass and pressure drop data are shown in Figures 18 through 20.
There was concern that conditioning may not be effective with an extremely high-ash fuel
such as the MJlner coal. The required conditioning agent concentrations were not known
and, theoretically, could be much greater because the inlet dust loading was almost 4 times
greater than with Big Brown coal. Nevertheless, starting concentrations were chosen at
37
Time (min)
Figure 17. Respirable mass emissions and pressure drop for baseline and conditioning tests
with Big Brown coal with woven glass fabric at an A/C ratio of 4 ft/min. Bags
were cleaned every 45 minutes after an initial 90-minute period.
38
102
i
10x ....................................... )._
100 ' i
• _ Iii _ i , ,,
104 -I! _ -_
' .......... _...i
......
i )
. .............. ) , _ .....
10"2.. •..... _............. _. .............. _ ......
g 10-3 i "
;....i............ .-). .............
10"4 i --
i
,o-, i i Milner C
; i ..
_
ImU
I............ _Te_:_:................
o _ I,..___ r...,...,,.._._,..__
0 100 200 300 400 $00
Time (min)
Figure 18. Respirable mass emissions and pressure drop as a function of time for Milner
• coal at an A/C ratio of 4 ft/min with Ryton fabric and no conditioning. After
an initial 90-minute f'fltration period, bags were cleaned every 30 minutes, except
• that the last two cleaning cycle intervals were 60 minutes.
39
°
103 ! i i i
102 .................. i¢"1 - -- .._

. .............. l_................ i...................
:
..................
_i - - ! i............
TM I................
i...................
11 _ l'<i>'i''i ' , !,, ,
..... ,...... "i';I.................. i.............
Ii
T;tl : i;
i , i
l O-t ........ !...... L_ I........... L..............
•i.
'_ "-'-'-"-
10° "' ..... '."".......
.....
' L ,'l
l!-'" `'
i >'L
("i_-' ........ 'ii .........
10-3 ..... ] --
! !
104
Milner Coal i
$ !_.yion.,F..ali_ i
i
_, AIC = 4 ft/min i
_ 3 --" :
'
0 l -- "
0 100 200 300 400 500
Time (min)
coal at an A/C ratio of 4 ft/min with Ryton fabric and 24)12 ppm of NI-I3/SO 3.
After an initial 90-minute filtration period, bags were cleaned every 30 minutes,
except that the last two cleaning cycle intervals were 60 minutes.
4O
103
10_
"_. 101 -.:.' ........................ _..._

E
100 , )
, : , _ I
__--_ 10"1 -'_

a..... i'-'( ............ I_..........t
"_" 10"2 i _--_I
10-3 ,_ _ ...... ..._..L.|

{
10.4
MilnerCoal
4 12 ppm _ ppm SOa
i_ AIC -- 4 ft/rain
r,.
0
0 100 200 300 400 $00
Time (rain)
• Figure 20. Respirable mass emissionsand pressure drop as a functionof time for Milner
coal at an A/C ratio of 4 ft/rain with Rytonfabric and 12/6 ppm of NH3/SO3.
• After an initial90-minutef'tltrationperiod, bags were cleanedevery 30 minutes,
except that the last two cleaning cycle intervals were 60 minutes.
41
24/12 ppm with the idea that they could be inc_ for the second test if necessary.
However, concentrations of 24/12 ppm were quite effective, so they were reduced to
12/6 ppm for the second conditioning test. Paniculate emissions for all three tests were very
low, achieving a collection efficiency of 99.99 %. Pressure drop and K2 were reduced by a
factor of three at concentrations of 24/12 ppm compared to the baseline test. At
concentrations of 12/6 ppm, pressure drop and K2 were slightly higher than at concentrations
of 24/12 ppm, but pressure drop was still considerably lower than with no conditioning.
Excellent results with 24/12 ppm of NH3/SO3 conditioning agents indicated that operation at
a higher A/C ratio should be possible, so a subsequent test was conducted at an A/C ratio of
8 ft/min (see Figure 21). Pressure drop was maintained at about 4 inches WC with a
20-minute cleaning interval, and particulate emissions were well controlled at 99.99 %
removal efficiency. Achieving superior performance at a high A/C ratio of 8 ft/rain with a
dust loading of 17.7 grains/scf again shows the significant benefit of using conditioning with
pulse-jet cleaning. Two additional tests, a baseline and a conditioning test at an A/C ratio of
4 ft/rain, were conducted with the MAlner coal using a P84/Ryton relied fabric in place of the
• Ryton fabric. Results, shown in Figures 22 and 23, were similar to the Rymn tests in that
pressure drop and K2 were reduced by about three times and paniculate emissions were low,
about 99.99 % collection efficiency in both cases.
5.3.3 _obesive Pro_rties and Ash Analyses
The effect of conditioning on ash properties is of intet_ to help understand the

mechanisms of performance improvement. The reduction of K2 must be attributed to
increased porosity of the dust cake because the panicle-size distribution of fly ash appears to
be unchanged with conditioning. While direct measurement of the dust cake porosity was
not attempted, aerated and packed porosity measurements of the baghouse hopper ash (given
in Table 6) are an indication of the packing tendency of the fly ash. From Table 6, one of
the effects of conditioning on the fly ash is a significant increase in both the aerated and
packed porosities, which explains the decrease in K 2. As discussed earlier, models, such as
the Carman-Kozeny relationship, which relate K2 to porosity, predict that a small change in
porosity can account for a significant change in K2.
The K: of the baseline Miiner ash is much lower than the Ks of the baseline Big Brown
ash even though the particle-size distributions are similar. Coulter counter results (presented
in Table 7) show the volumetric median diameter (VMD) of the Big Brown ash samples to be
14.8 + 1.6/_m and the VMD of the Milner ash samples to be 14.3 + 2.2 _tm. The
multicyclone results for the two coals also show the fly ash particle-size distributions smaller
42
• Figure 21. Respirable mass emissions and pressure drop as a function of time for Milner
coal at an A/C ratio of 8 ft/rain with Ryton fabric and 24/12 ppm of NH3/SO3.
• After an initial 40-minute f'fltration period, bags were cleaned every 20 minutes.
43
103 < _ _ i
i i i i
lO1 ) " '._"-_
...............q'O"
...._"
...... _ " _
_,oO i i i _ ..... -"""_ v_Fi
i 10.1 ,_....... _:--_-.-p, :...........

,o" I. ..'-'.,,.4-, t--
,o"
104 i
i
|
!
i
,i i
i
,:
. .
i
i
!
.:
iMilnerCc_I
i : o
i
:
, )i i P84/Rytoia Fabnc !
8 I : :-'-"---'_ '_ '
-'_ I ii_ i_A/C - 4 _mm

i !
•=- 6 • i
_
i
'.l .ii •
t_
•
i
#I i
!
|
:
i
|
' i ! ' q' )!

t _ _' ,1__,
0 ! () () i , i ,,()
0 100 200 300 400 $00
Time (rnin)
coal at an A/C ratio of 4 ft/min with P84/Ryton fabric and no conditioning.
After an initial 90-minute f'dtration period, bags were cleaned every 30 minutes.
44
. i _ ! i
102 ......................
._,_.................
Lt! 0 0 I.........................
i_ _ _ _ i..................,i
I ).! I z . - _ "r
" 101 •• "........
: ql I) "r......................
' m 4._-
.m" .--o.--.i
1] [r ....
I( J .
_: _
,o-,._..T
.... i'i
_ :'_,.:..i.'ooio,i
' ....
!"....
_._) i,__.
.....
.L_!L.h,.....,,..
lo.,----?
10_
,- i j---._.,
i i MilnerCdal i
5- i ) PS4/Syt 'o_Fabric )
! i24/)pm _ ' & 12 p_m

i SO 3
" 4" t _ _ )
e ) i
) ) ) i
i ! ) )
- ..... _---.-" _-" _.... F..... _ , )
_, ()))_,, )_,.)., .,__).!.a/)

' ,
0 100 200 300 400 500
Time (min)
• Figure 23. Respirable mass emissions and pressure drop as a function of time for Milner
coal at an AJC ratio of 4 ft/rain with P84/Ryton fabric and 24/12 ppm of
• N'H3/SO 3. After an initial 90-minute filtration period, bags were cleaned every
30 minutes.
45
than 10 _m to be similar (see Figure 24). Therefore, the porosity of the baseline Milner ash
must be greater to account for the lower K2. Measured porosities (shown in Table 6) show
that, indeed, this is the case. For an ash such as the Big Brown ash, that has a high
tendency to pack, an increase in porosity might be expected with conditioning. However, the
baseline porosity for the Milnor ash was considerably higher than the baseline Big Brown
porosity, and conditioning still produced a significant increase in porosity and a
correspondingdecrease in K2. This is an encouragingresult because it indicates that
conditioning is effective for ashes of varying properties--notonly in cases where the ash has
a high packing tendency.
Tensile strength measurements for the Big Brown and Miiner ash samples at an A/C
ratio of 4 f't/minwith Ryton fabric are shown in Figure 2.5. Tensile strengthmeasurements
for all of the fly ash samples for each coal are shown separately in Figures 26 and 27.
Looking first at the four Big Brown samples, shown in Figure 25, one can see thatthe effect
of conditioning is a shift of the tenSile strengthvs. porosity curve to the right and sometimes
upward. At concentrationsof 12/6 ppm of NH3/SO_,the curve shiftsprimarily to the right
and slightly upwardfrom the baseline curve. At concentrations of 24/12 ppm, the curve
shifts furtherto the right and upward. However, at concentrations of 50/25 ppm, the shift is
further to the right and ;,iownwardagain. The shift toward increasing porosity with
increasing concentra,Aonsis in agreement with the aerated and packed porosity
measurements. However, there is no obvious explanationfor the decrease in tensile strength
at the highest concenuation of 50/25 ppm of NI_SO_. The actual tensile mength of the
dust cake is of interest because increased tensile strength implies that the dust cake may not
be easily removed. Since measurement of tensile strengthrequires some compactionof the
ash, the tensile strength dam are generally for a porosity range that is lower thanthe actual
dust cake porosity. For these tests, dust cake porosities were not known, but, for previous
conditioning tests with Monticello fly ash, dust cake porosities were inferredbased on a
comparison of laboratoryIQ measurements with the dust cake K2. From the Monticello
tests, the tensile strength of the dust cake with conditioning was actually lower than the
baseline dust cake tensile strength. Therefore, even though tensile strength of the fly ash
may increase at constant porosity with conditioning, the actual dust cake tensile strength may
decrease with conditioning because of increased porosity. In addition, recent pore-bridging
tests in the laboratory show that superior pore bridging is achieved with conditioning even
though tensile strength of the conditioned ash is reduced under filtration conditions. An
explanation is that the particle-w-particle binding forces must be greater with conditioning.
46
: 90 ..... ' i ! iil
i a BigBrown i i i il
i z Milncr i i i il
70 ......_......................
i.............
"_ 50 _ __ _ : _ i ! !ii
30 ! _ ! ! i _ _ i
_9, !::iii _ i
gO ! i ! ::i i .:
on 10 : : _ : : _
i : : : : :
E !ii!::
..... _ .....
, i!ii!!
0.01 .....
1 10
Aerodynamic Diameter (.um)
Figure 24. Particle-size distribution of Big Brown and Milner fly ash based on multicyclone
results.
7 i n _o_,m _ppm
_z24
_. S03
-. 0 24 & 12 ppm
oJ_
om i
" 2
0
30 35 40 45 50 55 60 65 70
Porosity (%)
Figure 25. Tensile strength as a function of porosity for baseline and conditioned fly ash
• with Big Brown and Milner coals with Ryton fabric at an A/C ratio of 4 ft/rain.
47
I0
o
lm
s
6 I ......................
+................................
_ .....
t * "r,,++.. 2,,12,,,,m •
ooooooooo,o.+ooo .o.oo_ooooo.o.oo.oo_oooo_oooooooooo_o...ooo.o.oooo. oo,.,+o _koooo_ ,oo,
0
30 40 50 60 70
Porosity (%)
Figure 26. Tensile strengthas a function of porosity for Task 2 Big Brown tests.
mitr•'[ Coat 0 ' '

.Tem+10
tr '
dr.14 +. Bmeline [
Ci est 11 - 24/12 ppm I
+ ,_ i z tmr't+- 24/lZwm I I
l + • +,,i+.++,+,,,,.,,
N
+ + + +
_- t la ! i
= 3 ...................
'-".....
i i.......
: i.............
: _-...._.
:
c_ io i ,, !
2 ..........................
_..................
_ ..............._ .......
i ....._...........
= 1 .......................... °°
i oo i i °°"'°°°_°
11.i ii i v i l i i i i
50 55 60 65 70 75
Porosity (%)
Figure 27. Tensile strength as a function of porosity for Task 2 Milner tests.
48
The tensile strength curves for the Milner samples (Figure 25) shift toward higher
tensile strength and somewhat toward higher porosity with conditioning. However, since the
porosity was higher with conditioning, the data imply that the tensile strengthof the
conditioned dust cake was no greater than the baseline dust cake. From Figures 26 and 27,
it is evident that there is some variability in the tensile strength data. For the Big Brown
results, one would expect the four ash samples collected from tests at concentrationsof
• 24/12 ppm of NH3/SO3to have similar tensile strengths. However, Figure 26 shows a wide
range in the tensile strength values at the highest compaction level (e.g., at 50% porosity).
Similarly, Figure 27 shows some variability for the three Milner samples at concentrationsof
24/12 ppm of NI-I3/SO3. The reason for this data spread is not clear. From previous bench-
scale tests conducted for DOE, relative humidity is known to have an effect on tensile
strength, especially for conditioned samples. To minimiTe this effect, samplesare stored for
an extended period at 10% relative humidity prior to the tensile strength measurement.
However, there may be some differences in the time the samples axe exposed to the ambient
laboratory conditions during the measurementprocess, which could accountfor some of the
variability. To obtain more quantitativedata may requireconducting the tensile strength tests
completely within a controlled humidity chamber. Nevertheless, the tensile strength data
show that conditioning produces a shift in the tensile strength vs. porosity curve toward
higher porosities and sometimes toward higher tensile strength. However, extrapolatingthe
data to the higher porosities of a dust cake implies that the dust cake tensile strengthdoes not
necessarily increase with conditioning, which alleviates concern about bag eleanability
because of possible increased tensile strength.
Major elemental concentrations of the baghouse hopper samples measured by x-ray

fluorescence analysis are given in Table 7. These data indicate thatconcentrations of major
elements were fairly constant from test to test, as would be expected. A small increase is
expected in SO3 because of the amount of SO3injected. For example, injectionof 25 ppm of
SO3 with an inlet dust loading of 4.55 grains/scf should increase the fly ash SO3content by
0.81% SO3. Injection of 12 ppm of SO3should result in a 0.39% SO3increase. While the
data in Table 7 do indicate an increase in SO3 on the order expected, there is some
variability in the SO3 concentration for the baseline tests. Therefore, a quantitative
determination of the amount of injected SO3that is transferred to the fly ash is nor available.
Similarly, for the Milner tests, since the average dust loading was 17.55 grains/scf, 12 ppm
of SO3 should produce an increase of 0.10% SO3 in the fly ash, which does not allow for a
quantitative determination of SO3transfer.
49
Two ammonia slip measurementswere conductedfor each conditioning test as part of
the EPA-Method5 dust loadings. For the Big Brown tests, the indicated ammonia
concentrationin the flue gas ranged from 0.9 to 5.1 ppm with a mean value of 3.0 ppm and
a standarddeviation of 1.2 ppm. This indicates that most of the ammoniawas transferredto
the fly ash and that ammonia slip is not a problem. For the Miiner tests, the measured
ammonia level ranged from 0.1 to 0.7 ppm, with a mean value of 0.3 ppm and a standard
deviation of 0.2 ppm. The lower ammonia slip comparedto the Big Brown may be the
result of the much higher dust loading with the Milner coal which means that there is much
more available surface area for the ammonia to be adsorbedon the fly ash. Ammoniain the
fly ash was also measured, but the appment amountof the injected ammoniathat was
transferredto the fly ash ranged from only 30% to 75_. A combinationof the ammoniaslip
and the ammonia transferredto the fly ash should accountfor ali of the injected ammonia.
However, the closure was low and the unaccounted-forammoniaranged from 15% up to
70 %. The ammonia slip values appearto be reasonable, and conditioning hada significant
effect on baghouse performance, which implies that the conditioning agents must have been
transferredto the fly ash. This points to the fly ash ammoniaanalysis as the source of poor
closure. The procedurefor separatingthe ammoniafrom the fly ash may have resulted in
some off-gassing of ammonia, or possibly the ammoniawas not entirely e_, eitherof
which could account for lower ammonia concentrationsthan expected. The e_on
procedure was furtherevaluat_ for the Task 3 te_ in an attempt to improvethe closure for
the ammonia balance.
5.4 Task 3 Results (100-hour and 16-ft/rain tests)
A summary of the test variables and results from the Task 3 tests are shown in Tables
8 and 9.
5.4. I 100-hotlr T_sts with Bi_ Brown Coal
Three 100-hour tests were completed with Big Brown coal in Task 3. These included a
baseline test (Run 403) at an A/C ratio of 4 ft/min and conditioning tests at A/C ratios of 4
(Run 404) and 6 ft/rain (Run 405). Ali three of these tests employed Ryton fabric, and
concentrations for the conditioning tests were 24/12 ppm of NH3/SO3,as shown in the test
matrix presented in Table 2.
Plots of _P as a function of time for the three 100-hour Big Brown runs are shown in
Figure 28. The bags were pulsed once per hour in each of these tests. Whenever bag
50
51
52
8 K
i ...... i i
I ..... _e]_e (_i: = 4w_,)""

J/ Before & After Cleaning -"--
-----" Conditioned
Conditioned (A/C : 46 ft/rain)
(AIC ffi ft/min) ,
..... • ...o.o 0t 1_ t
•.'"i"
_e4 6 , *,,L'T
............................................................... t..........
.," a...........
0''I,_ .'y*.'-
:it
-- e_ ft • -_ 0 Lt e --:,_"
'k" " ...........
b D
mm /I , ,,:: ,'
., .::, -,,,• -
• • II til
'I :":" ...............................................

_ 2 ....................... ".........................
0 20 40 60 80 100 120
Time (hrs)
Figure 28. Effect of conditioningon baghousepressure drop for 100-hourBig Brown tests
for Runs 403, 404, and 405. Bags were pulsed once per hour.
cleaning
occurred,
allofthebagswerecleaned
inrapidsuccession
(within
6 seconds).
Therefore,
theZkPbeforebagcleaning isthembe sheet pressure
dropaftera I-hourperiod
ofdustcakebuildupon allofthebags.The Z_Pafter bagcleaningisthetubesheetpressure
dropimmediatelyafter
cleaningallofthebagsandisduetotheresidual duston thebags.
Forthebaselinetest
atan A/C ratioof4 ft/rain,
thebefore-cleaning
Z_Pclimbed steadily
from2 inchesWC atthestartofthetest to5.5inches WC by theendofthetest. The after-
cleaning
Ap alsoclimbedsteadily
fromnearzeroatthestart toaboutI inchWC by theend
ofthetest.FortheconditionedtestatanA/C ratio of4 ft/min,
thebefore-cleaning
zIP
reachedonly1.7inchesWC, andtheaftor-clem_ingAp was only0.3inches WC. Twenty
hoursintothetest,
theAp was verysteady, whichisan encouragingresult
becauseit
indicates
theresidual
dustcakewas notbuildingup excessively
asa result
ofthe
conditioning.
The initial
andeffectiveK2 valuesforthesetests,
showninTable8,area
• factor
ofthreelowerfortheconditioningtestscomparedtothebaselinetest.Thisisalso
shown in the pressure drop data, which indicate the Ap for the baseline test was about three
• times higher than the z_Pfor the conditioning test when the A/C ratio for both tests was
4 ft/min.
53
'°,
For the condition_ test at an A/C ratio of 6 ft/rain (Run 405), the before-cleaningAP
increased for the first 30 hours and then remained steady. However, 50 hours into the test,
the aP decreased and then was f_Lirlyconstant again for the remainder of the test. There
is no apparent reason for the decrease in AP at 50 hours, but it could have been caused by a
small shift in the fly ash particle-size distributionand/or a decrease in the dust loading.
Since K2 is inversely proportionalto the square of the particle diameter(see F.quation2), a
small change in diameter(for example, from 14 to 12 _m) could increase I_ by 35%. With
a constant bag-cleaning interval and a constantK2, the AP at an A/C ratio of 6 ft/rain should
be 2.25 times greater than the AP at an MC ratio of 4 ft/rain. Since the actual AP at
6 f't/min was only about 1.5 times greater than the _LPat 4 ft/min, the effective K2 values
must have been different. The actual measuredeffective K2 value for.Run 404 was 3.6,
while the effective K2for Run 405 was 2.4 inches of water-ft-min/lb. Whether this
difference was the result of a small change in particle size is not known. Nevertheless, for
both conditioned runs, initial and effective Ks values were much less than the K2values for
the baseline run, which demonstratesthe benefit of using conditioning at both MC ratios.
Particulate emissions for these 100-hour tests with Big Brown coal are shown in
Table 8 and in Figures 29 through 35. The effect of conditioning is best shown in Figure
29, which gives the integratedaverage respirablemass emissions as a functionof run tiwe.
These values representthe average respirable mass emissions obtained from samplingduring
an entire l-hour filtrationcycle. The actual range in respirablemass emissions that occuned
for each filWation cycle is shown in Figures 30 through32. From Figure 29, there is tittle
difference in emissions between the baseline and conditioned tests during the first 8 hours of
the test, which is consistent with results from the 8-hour tests conductedin Task 2 with the
Ryton fabric. However, by 80 hours into the tests, Figure 29 shows that respirablemass
particulate emissions are two to three orders of magnitudelower with conditioning. The
EPA Method 5 data (see Table 8) also show emissions to be much lower with conditioning,
although the difference is only one to two orders of magnitude. The lowest dust loading
measurements were 0.0001 grain/scf, which is the limit of sensitivity that can be achieved
with EPA Method 5, unless very long sampling times are employed.
Both the respirable mass and EPA Method 5 data indicate that particulateemissions
were lower for the conditioned test at an MC ratio of 6 ft/rain than for the conditioned test
at an A/C ratio of 4 ft/rain. This might appearcontrary to expectations, because, typically,
emissions increase with an increase in MC ratio. The difference, however, can be explained
if one considers that both tests were conducted with a constant cleaning interval ratherthan at
constant pressure drop. At the higher A/C ratio, the pores are more rapidly bridged after
54
': 102:
4mm
Big Brown Texas Lignite ..........
area
" imm
101
I • Baseline ) i
] 0 24ppm NH3 &=412pp
., ......o,..o..... ft/rain (AIC
o....,.......,..
,..o..........
_."_._°_._._.._._._._°_°_°_._"_%_°.._-_°_°_"".-_._._..
-" m SO3 ffi4 ft/mm) !

• , mm q
mim
O 1 _x24 ppmNH3& 12ppmSO_(AIC= 6 ft/rain) I
e_ A
.......................... ..................,......_......
10o ,_.
mm
lm
• _ 10"1
e_ ,_ l0 "2
10.3
0 20 40 60 80 100
Time _rs)
Figure 29. Effect of conditioning on average respirable mass emissions for 100-hour Big
Brown tests for Runs 403, 404, and 405.
103 ! ....
_ Big Brown Texas Lignite

_ 101 'KTC"_"_'ttTii_
= Baseline
C
"_ I0° J O AfterBag Cleaning ! ....[
E ] .......
_,l 10 "x
r_
"_- • Priorto BagCleaning[ !
10.2 .
0
,__ 10-3 .....Q ....................................................................................................
.9..........................
t_.
emm $
| || ......................... ° ........... ,...., ............... °° ..... °..... °............ . ................ °o°°°, .... ._**... ..... ..oo.o._._...oo.°o°.
.5
10 " ' I " ' ,' I ' ' I
0 20 40 60 80 100
Time (hfs)
• Figure 30. Respirable mass emissions before and after bag cleaning for the 100-hour
baseline Run 403 with Big Brown coal.
55
' 103 ! BigBro'wn Texas'Lignite
A/C = 4 ft/rnin
_.
_MU 1o: --2_____-r2-p_so_ .....................................
101
.i
._ I00
10.2 I • .........
Prior to BagCleanmg _ [
I
3 10.3 ..............................................................................
"_' Q ,. ,
10.4 ................... • t S,....
[
I
...................................................
,,
.......................
,.....................................
10"s , , • , @ , ' I
0 20 40 60 80 100
Time (hrs)
Figure 3 I. Respirable mass emissions before and after bag cleaning for the 100-hour
• conditioning Run 404 with Big Brown coal.
10 3
= 6 tr/rain
10= 3
10x
rd_
o 0
•-_
r_
10 0
(l,mm
u_
10"2 [ O AfterBag Cleaning

•-_))11_ I
_..*--,::::--'_''*'*'*'****'*'*''*''**'**'**'''*'**'_'''
)_ | o. Prior to Bag Cleamng [

10"3 ,..o............................. _ .............................................................
._.
tlo
o5
I0 ' I ' , ' ' ' ' '
0 20 40 60 80 100
Time _rs)
Figure 32. Respirable mass emissions before and after bag cleaning for the 100-hour
conditioning Run 405 with Big Brown coal.
56
". 10s .:' Big Brown Texas Lignite
e_
_0'_................................
.o............................
,9_
............
_._]!:.__..............
104 " "_-_" ...........................

"_"................................... 4_ .......... 0
• "_ . _......................................................................................
_ .....................
xo_ _:_ ___
l0 s -_-_ .......... O0 ......................
_ ..........................
._ lOt _0 ...................
_ ............................
0 ................. o ....
_.- 10°
lO.t ._
.............................................................
6 ...................
-- ........................
_ ........................
_
........
.................m_C
I
10.2 , j ' i ' , ' 4 .... ' /
0 20 40 60 80 100
Time(hfs)
Figure 33. Range of submicron particle emissions for the 100-hour baseline Run 403 with
• Big Brown coal.
108 ................
_-, 10_ AVL"_--4"tWmib

!a_
'_
lOs
104
103
---------
_0'-.._.......................
-_.0
_ ...........................................................................
_---
-_...
i
1o.,I.
................
100
_..............
10.2 _ ...........
0
:.........
..................................................................
'
20
, ' ,
40
'
.................. 0 .................
60
I '
80
J '
100
Time (hrs)
• Figure 34. Range of submicron particle emissions for the 100-hour conditioning Run 404
with Big Brown coal.
57
lOs
24 ppm NH3& 12ppmSO3 Big Brown Texas Lignite
A 10_ TA-/U'"_"'6-t_i" ...........................................................
.. lOs
-_ 104 -- ,
10o LO0 0 .......

10"s
104
0 20 40 60 80 100
Tune (hrs)
Figure 35. Range of submicronpanicle emissions for the 100-hour conditioning Run 405
with Big Brown coal.
bag cleaning, which means there is a smaller percentage of time between cleanings with open
pores (when most of the emissions occur). Therefore, the net result was lower emissions for
the higher A/C ratio. However, ff the bags were cleaned on a _P basis ratherthan on a time
basis, emissions would likely have been aboutthe same at both A/C ratios.
The range in submicronparticle count, measured with the CNC, is shown in Figures
33 through 35. The CNC actually read zero counts/cre3 for many of the lower readings, but
these are reported as 0.08 particles/cre 3, because the sensitivity limit of the CNC is 0.01
particle/cre 3. Since the sample delivered to the CNC is diluted 8:1 with particle-free air, ali
CNC readings were multiplied by a factor of 8 to obtain the actua._,particle concentration in
the flue gas. The submicron emissions ranged from about 0.08 to 106particles/cm3 for both
the baseline and conditioning tests. Since the CNC detects particles over the entire range
from 0.01 to 1.0 _m, low CNC readings provide strong evidence that ali submicron particle
emissions are very low. At the lowest reading of 0.08 particles/cm3, complete pore bridging
has occurred, and the submicron collection efficiency is _e,ater than 99.9999%. The
58
emissions
thatoccurwhen there
iscompleteporebridging
areprobablyduetogradual
seepageofdustthroughthefabric
andsubsequent
reentrainment
ofdustfromtheclean side
ofthefabric.When dustisreentrainexl,
submicronparticles
willgenerally
notbe recntrained
asindividual
particles.
Datashowthat,priortobagcleaning,
submicron emissions
were
nearzeroforboththebaseline
andconditioned
tests,
buttherespirable
massemissions(from
I toI0_m) wereclearlyhigher
forthebaseline
test.Thismdica¢cs
thatsignificant
• reentralnment
occurredforthebaseline
test,butnotfortheconditioning
tests.
The pressure drop and particulate emissions data demonstrate the benefit of
conditioning with Big Brown coal. Pressure drop and the correspondingKz values were
reduced by a factor of about three, and particulate emissions were reduced by two orders of
magnitude. The EPA Method 5 measurements indicate that particulate emissions for the
baseline test were 0.018 11)/106Btu, which does not provide much of a safety margin from
the regulatory emission limit of 0.03 11)/106Btu. In addition, for the baseline case, there is
little confidence that the A/C ratio could be increased beyond 4 ft/min, without exceeding the
regulatory limit. However, with conditioning, particulate emissions were substantially
reduced for both A/C ratios tested (0.0005 to 0.0014 lb/106 Btu), and the At) was lower with
conditioning at an A/C ratio of 6 ft/rain than the Al' for the baseline test at an A/C ratio of
4 ft/min. The explanation for reduced particulate emissions and pressure drop, which are
discussed later in this report, involves changes in dust cha_._ristics as a result of
conditioning.
5.4.2 100-hour Tests with Milner Coal
Three 100-hour tests were completed with Miiner coal in Task 3: a baseline teat
(Run 406) at an A/C ratio of 4 ft/rain and conditioning tests at A/C ratios of 4 ('Run407) and
6 ft/min ('Run 411). AI1three of these tests employed P84/Ryton fabric, and concentrations
for the conditioning tests were 24/12 ppm of NH3/SO3,as shown in the test matrix (Table 2).
Plots of _ as a function of time for the three 100-hour Milner runs are shown in
Figure 36. The bags were pulsed once every 30 minutes in each of these tests. Because of
the very high ash content, the combustor had to be periodically shut down for ash cleanout,
which resulted in about 70 hours of actual f'fltrationtime for Runs 406 and 407. For
Run 411, almost 90 hours of filtration time were achieved by employing a different
combustor bottom, which allowed for ash removal from the combustor without shutting
down. The AP data shown in Figure 36 are presented as continuous plots, ignoring the off-
Lineperiods. Comparing the At' plots between the baseline and conditioned tests at an A/C
59
6
' Milner Coal Before & After Ck,m_
= 4 . ,.-. ::,i , I , n M
,.gels# . _, , ,e .J, ... . .,. ._'. o .
...: .. JL'.' j • . : .e "IA: i_
J'_
_:
0 20 4O 6O SO I00
Tune Cnrs)
Figure 36 Effect of conditioning on baghouse pressure drop for 100-hourMiiner tests for
Runs 406, 407, and411. Bags were pulsed every 30 minutes.
ratio of 4 ft/rain, the baseline ALPis about 2.5 times greater thanthe _P for the conditioned
test. Initial and effective K2values (from Table 8) are 2.3 and 2.4 times higher,
respectively, for the baseline test. For the conditioned test at 6 ft/rain (Run411), the
effective K2 value was about 2.2 times lower than the effective K2value for the baseline test.
A combination of an increase in A/C ratio from 4 to 6 fOmin and a decrease in K2by a
factor of 2.25 should result in identical _P plots for a constant dustloading and cleaning
cycle. Figure 36 reveals that the Ad) with conditioning at 6 ft/rainwas slightly less than the
baseline Ad_ at 4 ft/min. However, a somewhat lower dust loading for the conditioned test at
6 ft/min (see Table 8) accounts for the slightly lower ALP. All of the _t ) plots indicate a °
gradual trend of increasing _P with time, which indicates that the tests were not long enough
to establish steady-state operation.
Paniculate emissions for the 100-hour tests with Milner coalare shown in Table 8 and
in Figures 37 through 40. Figure 37, which gives the integrated average respirable mass
emissions as a function of run time, shows that, by the end of the tests, particulateemissions
were almost two orders of magnitude lower with conditioning. The EPA Method 5 dust-
loading data also indicate that particulate emissions were significantly lower with
conditioning. For the baseline case, emissions were already low at 0.0059 lb/106 Btu, a
60
• 102
•, _ Milner Coal
=.
_1) ConditimmdTram,- 24 ppm NH3 & 12 ppmSO3
101 .......................................
"_ 0 IIueline (A/C"4 h/rain) I
@ Condtioned (AIC=4 f',./min) I
.2 _ Condt/oned (A/C=6 f'dmin)
,sm
100 ....... J............
drum
•
10-1
1,)
,_ 10-2
10.3
0 20 40 60 80 I00
Time (hrs)
Figure 37. Effect of conditioningon averagerespirablemassemissionsfor 100-hourMilner

tests for Runs 406, 407, and 411.
103 ! ....
E 10: I_ _ _._
•_ 10_ _ @ O"
_mm
mm
O
•_. _o° ...._e_-c_ FO A_rSagCie_i.s i

"_ 10"1 --_'_IiheA/C
= 4 ft/rain I • Prior to Bag Cleaning LI
•
10-2
"e"• • ._
I
..........................
-o
10-4 ..
10"s
0 20 40 60 80 100
t,
Time (hrs)
Figure 38. Respirable mass emissions before and after bag cleaning for the 100-hour baseline
Run 406 with Milner coal.
61
'.
10 3 _ .....
102
_ 101
) J
o After]3agCleaning
Priorto BagCleaning
•2 0
el
N lO ,
_ lO --'I_T_'C_"aI
A/C ffi4 R/min
:_ 10"_-'2_-p-__--_12W_so_
10"3 "_ • *
.._ • •
10.4 ...............................
i0 -s
0 20 40 60 80 100
Tune (hrs)
conditioning Run 407 with Milner coal.
103
A/C = 6 R/rain
"_ i02
3
,_ 10x
O
o
t_mm
100
omm
N 104
o 10"3i *
°
Priorto Bag Cleaning[
_- 10"4
i0-5
0 20 40 60 80 100
Time(hfs)
conditioningRun 411 withMilnercoal.
62
factor of five times better than the regulatory limit of 0.03 lb/106 Btu. The range in
respirable mass measurementsprior to bag cleaning to just after bag cleaning for these three
runs is shown in Figures 38 through 40. The submicronemissions data, shown in Figures
41 through 43, indicate that the peak submicron emissions were higher with conditioning.
This is not surprising, considering that the peak submicron emissions were only about
1000 particles/cre 3 for the baseline test. In cases where the baseline submicron particle count
is low, the effect of conditioning will generally be to increase the submicron particle
concentration because of the formation of some fine ammonium sulfate particles. If the bags
are cleaned very well and there is only a light residual dust cake, the submicron emissions
spike that occurs after bag cleaning may be higher with conditioning. However, conditioning
generally facilitates rapid pore bridging, so this sur;micron emissions spike is short-lived, and
the total submicron emissions are not a concern.
The pressure drop and particulateemissions data show that conditioningworks well
with Milner coal. K2 values were reduced by a factor of 2 to 2.4, and particulateemissions
l0 T .................
Milner Coal
10 6. :_../C..-_.--4.Wfffffi
lOs Baseline
104 o
w 103 0 ..... _ .....
101 O
e_
100 OQ
" 10
m 10-2
10"3
0 20 40 60 80 100
i,
Time (hrs)
• Figure 41. Range of submicron particle emissions for the 100-hour baseline Run 406 with
Milner coal.
63
. 107 24 ppm NI-Ia & 12 ppm SO3
106 = 4 ft/m.'_
104 ...._P..
........... _ ._ .....
102
o
VO'-
............
, _"
.[ 10° ---
._ i0 "l
m 10-2
10.3 , _ , , , , ' l '

0 20 40 60 80 I00
Tune (hrs)
Figure 42. Range of submicron panicle emissions for the 100-hour conditioning Run 407
with M.ilner coal.
los I M/lner Coal
'_
10*/__[
106
24ppm_ & 12ppmSO3
_ 10$_ "- _" '_-: ,_ ..
E 1°_
= 10.x .
Or_ 10.2 -_............................................................................................................................................

10.3 " , , ' _ ' l ' , '
0 20 40 60 80 100
Time (hrs)
Figure 43. Range of submicron panicle emissions for the 100-hour conditioning Run 411
with Milner coal.
64
were reduced one to two orders of magnitude. Ash properties for the Milner tests are
discussed later in this report.
5.4.3 100-hourTests with Black ThunderCoa.l
The original test plan included only two 100-hour tests with Black Thunder coal: a
. baseline test at an A/C ratio of 4 ft/rain and a conditioning test at an A/C ratio of 6 ft/rain.
It was thought that, if conditioning worked well for the other two coals at an A/C ratio of
6 ft/rain, it would also work well for the Black Thunder at an A/C ratio of 6 ft/rain. The
baseline test (Run 409) was conducted without major difficulties. However, after less than
24 hours into the conditioning test (Pain 410), it was obvious that the baghouse &P could not
be adequately controlled, so the run was stopped. The reasons for the difficulty in bag
cleaning were thought to be related to either the high conditioning agent concentrations
(compared to the relatively low dust loading) or to the high A/C ratio. There was also the
possibility that conditioning would simply not work with this coal. To address both of these
)
possibilities, new bags were installed, the MC ratio was reducedfrom 6 to 4 ft/min, and the
conditioning agent concentrationswere reduced from 24/12 ppm of NHs/$O3to
12/6 ppm of NHs/$O3 for the remainderof the 100-hour test. The first pan of the 100-hour
test at an A/C ratio of 6 ft/min was designated as Run410A, and the last pan of the test at
an A/C ratio of 4 ft/rain was designated as Run 410B. Since conditioning workedvery weLl
for Run 410B, an additionaltest was conducted at an MC ratio of 6 ft/min, but with
conditioning agent concentrations of 12/6 ppm of NTIs/SOs(Run 413). This last test was
possible because Run 412 was terminated early as a result of pressure drop problems.
A comparison of the _LPplots for the three Black Thundertests is shown in Figure 44.
Comparing the baseline (Run409) and conditioning (Run41OB)tests at an A/C ratio of
4 ft/rain, the _ for the baseline test is two to three times higher, which is also reflected in
the K_ values shown in Table 8. For both tests, there is a trendof increasing _J' with time,
but the baseline _d_ appearsto be increasing more rapidly than the Ap with conditioning.
Again, this is encouraging because, within the limits of the test period, no problemsof
residual dust cake buildup or bag cleaning are indicated. These results show that
conditioning can be employed with Black Thunder coal to achieve a substantial reduction in
pressure drop at a constant A/C ratio. When the A/C ratio was increased to 6 ft/rain with
• conditioning (Run 413), there was a greater increase in zkt'than the factor of 2.25 times
increase expected from the increase in face velocity. The additional zkPincrease is reflected
• in the higher K: values for the test at 6 ft/min. Since no baseline test was conducted at
6 ft/mm, a direct determination of the benefit of conditioning at the higher MC ratio cannot be
65
0
0 20 40 60 80 100
Tune(hrs)
Figure 44. Effect of conditioningon baghouse pressure drop for 100-hour Black Thunder
tests for Runs 409, 410B, and 413. Bags were pulsed once every hour•
made. However, the K2 values with conditioningat 6 f't/minwere still much lower than the
K2 values for the baseline test at 4 ft/rain. The difference in K2 between baseline and
conditioning tests, both conducw.Aat 6 ft/rain, would likely have been greater, since K2tends
to increase with increasing velocity.
A plot of the _ for Run 410A (the first attemptedconditioningtest at an A/C ratio of
6 ft/rain with the higher conditioning agent concentrationsof 24/12 gnn of NH_SO3), shown
in Figure 45, indicates the problem that can occur when increasing the A/C ratio. The plot
shows that z_Pwas rising rapidly. To help control _P, at 8 hours into the test, the bag-
cleaning frequency was increased from once per hour to once every 30 minutes. However,
the _P continued to increase, which corresponded to a buildupin the residual dust cake, as
seen though the sight ports. At 12 hours into the test, the bags were pulsed off-line seven
times to remove the residual dust cake, and the test was continued. However, the _P again
rose uncontrollably, so the test was terminated. No attemptwas made to increase the pulse
pressure or pulse duration to control pressure drop. These results indicated that conditioning
with concentrations of 24/12 ppm would not provide satisfactory results at this A/C ratio.
The clam from Run 413, with NH3/S03 concentrations of 12/6 ppm, indicated much better
66
• 12 | Black Thunder Sub'J_iLuminous*C0al ......
24 ppm NH3 & 12 ppm SO3
10
I "--- kfor_C'u_n I
I...... I P
Chtniled
_. czunlnz
_ /
" -_ cyo,o/ .
= 6 ..............................................................
"_'_'f
..........
_:.................................
7'.........
// "" ::__ Lbalhou_
I cleaned J /
_ e
t I
t eeB o
o"
• ,i °
• t •
:olo oo ° •
eeooeoeoeoe41°° to °°
-0"0 i ii iii i
- e_ea_ ' e ® ° e'D'w I ' _ I °° o _ H_ ' _ I a e_°° '
0 5 10 13 20
Tune(hrs)
Figure 45. Effect of conditioning on baghouse pressure drop for Black Thunder test with
24/12 ppm of NH3/SO3 for Run 410A. Bags were pulsed once per hour for the
first 8 hours and once every 30 minutes for the remainder of the test.
results, but it appears that operating at an A/C ratio of 6 ft/rain with the Black Thunder coal
is difficult. This is somewhat surprising because the inlet dust loading was lowest with the
Black Thunder coal. Discussion of dust properties (later in this report)will provide some
insight into possible causes for this bag-cleaning difficulty, but the exact reasons are not
clear.
Average respirable mass particulate emissions, shown in Figure 46, and the dust-
loading data from Table 8 both indicate very little effect of conditioning. However,
emissions for both the baseline and conditioning tests were at least 35 times lower than the
regulatory limit of 0.03 lb/10 6Btu. Respirable mass emissions for each of the tests are
shown in Fig-arcs 47 through 49, and submicron particle counts are shown in Figures 50
through 53. From these graphs, it appears that, at an AJC ratio of 6 f't/min, both the
respirable mass and submicron emissions did not reach as low a level before bag cleaning as
• did the emissions for the conditioning test at an AJC ratio of 4 ft/rain. This may indicate
some deterioration in performance with an increase in A/C ratio from 4 to 6 ft/rain, which is
• in contrast to the results for the Big Brown and Miiner tests. They indicated no deterioration
in particulate removal efficiency as a result of increased face velocity. When bag-cleaning
67
,, 102 ........................
,._
col " Black Thunder Subbituminous Coal
U
,_ 101
ConditioMd Tssts - 12 ppm NH3 & 6 ppm SO3
_
lOO ! • '
10-,!!
_ . _- .
m 10"2 . ..................................... .,..........................
10.3 ' l" ' , ' , ' "'l " '

0 20 40 60 80 100
Time Oars)
Figure 46. Effect of conditioning on average respirable mass emissions for 100-hourBlack
Thundertests for Runs 409, 41013,and 413.
103 i"" Black Thunder $_ibbitu_ous c0al

AJc = 4/t/rain
lOZ ......
B_line
U
101 ............ @ : ..... ___
lo° o
10.2 -_. ................

@ •
mms
•_e_ 10-3
'--
o amunl • • oo db •
_--
10..4 .......9.........................................................................................................................
10-s ' '' I ' ' l , ' , l' ' I '
0 20 40 60 80 100
Time (hrs)
baseline Run 409 with Black Thunder coal.
68
.. 103 '_ Black Thunder Subbituminous' C0al ' ' -
A/C = 4 ft/rain
$_ .........................
0
101 .........................................
"_ 10° "_.................... o o
• _ 10" 0 After Bag Cleaning I.......

.
"- ._iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii iiii
10-3
Prior to Bag Cleaning!
.ffi 10
_'_ 10-4 ..._...................................

• ...........................................
O
10"s
0 20 40 60 80
Time (hrs)
Figure 48. Respirable massemissionsbefore and after bag cleaning for the 100-hour
conditioning Run 410B with Black Thunder coal.
lOS BlackThunderSubbituminous
Coal....
,'_' A/C ffi 6 R/rain
102 .......
I2ppm=_ _ 6'ppmso 3
10_ _.,---__
-- _
_0 e,_.,O
©
omm
lO0
omm
=_
lOq J 0 After Bag Cleaning I_
I • Prior to Bag Cleaning
I
lO.Z .........
-_ 10.3 .....................
o_
¢C
t_
e_m
1 O ........................................... ".......... ""°"'"'"''"°" "'"'°"_ _°'" ....... "'°°"°_°"_" ......... °'°"°_ ..... °_'
I0"s , _ ,,,,_ , _ , _ , j , ,j ,,
0 10 20 30 40 50 60 70
Time (hrs)
. Figure 49. Respirable mass emissions before and after bag cleaning for the 100-hour
conditioning Run 413 with Black Thunder coal.
69
t t
107
| Black Thunder Subbituminous Coal [
_. ......ALC_.=._4..fl/min I
106 ! Baseline
104 -- - • --- o--
_c 1o' o........ _%
O _
102 ......... .o.... _ ................ _ ........................
101 0 ...................
.__ 10°
j 10"1
10-2 ..
10-3 , , I , ,
0 20 40 60 80 104)
Tune(hrs)
Figure 50. Range of submicron particle emissions for the 100-hour baseline Run 409 with
' Black Thunder coal.
10s Black Thunder Subbituminous Coal

^ 10 7
'_=. i2 ppm NH3 & 6 ppm SO3
t_ 10_
l0 s
Z
'_ 4
10
-_ 103 o 0
lo_ ab
=
t...
lo_ o ..........
_ qe --_r--
•_ 100 ,. -_ (_ "q_--"
.._ - 10 "1 ........................... _ .....
10-2
10 .3
0 20 40 60 80
Time (hrs)
Figure 51. Range of submicron particle emissions for the 100-hour conditioning Run 41OB
with Black Thunder coal.
7O
' 10S '| Black Thunder Subbimmin0us Coal
_..A/.C_=....6.._mb.
.........................................
_" 107 _ 12 ppm NI-I3 & 6 ppm SOs
104 1.....................................................
OC
_ 1 .._.....
..............................
o...........................
"(30
_ ........................
7-02 ...................... .......................

•• "_
=L.
0
103
101 1 ..............
........ _._Am," =_.__
..................... ___
°o
1ill
,o,
o, 0 2O 40 60 80
• Time (hfs)
Figure 52. Range of submicron panicle emissions for the 1DO-hourconditioning Run 413
• with Black Thunder coal.
e_
102 BIaC-"hunder Subbituminous Coal
24 ppm NI-I3 & 12 ppm SO3
101
lm
g ,
10° -=-_--_
olm 10"x
10.3 _---_
t..
"=_ 104
10"s ,
0 2 4 6 8 10
" Time (hrs)
. Figure 53. Respirable mass emissions for Black Thunder test witl_ 24/12 ppm of NI"I3/SO3,
Run 41 OA.
71
problems occurred in Run 410A, particulateemissions remained low, as shown in Figures 53
and 54 and Table 8. Therefore, control of baghouse pressure drop appearsto be more of a
concern than particulate emissions for the Black Thundercoal using Ryton fabric.
5.4.4 16-ft/m_ Tests with Big Brown Coal
Five 8-hour tests at an A/C ratio of 16 ft/rain using Ryton fabric were completed. The
tests included baseline and conditioning runs with a normal dust loading and baseline and +
conditioning runs with a reduced dust loading. To achieve a reduced dust loading, the
combustor was fired on naturalgas while coal was fed at approximately1/20th of the normal
feed rate. Conditioning agent concentrationsfor the normal dust-loadingconditioningtest
were 24/12 ppm of NH3/SO3,and for the reduced dust-loading conditioningtests,
concentrations were 6/3 ppm of NH3/SO3. One other major difference for these tests was
off-line pulsing, in contrastto en-line pulsing employed for the other tests.
For the normal dust-loading baseline test (Rna 399), the single 10-ftbag was cleaned
every 10 minutes at the startof the mn, but by 2.5 hours into the test, the cleaning interval
i0 s ,,
Black Thunder Subbituminous Coal
_, 10_ -_' ........... 24ppmNH_ & i2ppmSo 3...... +
,,
,o,. .....
_ L ' -
_=
10 '41
103
_I+_
"''ml, +
..... "" ' --
• t , '_ t
.+,o,
__
ml
10° ..- ..... ) _ ....
c_ lO-I
10-2
0 2 4 6 8 10
Thne 0n's)
Figure 54. Range of submicron particle emissions for Black Thunder test with 24/12 ppm of
NH3/S03, Run 4 IOA.
72
had to be reduced to 5 minutes to control the pressure drop under 10 inches WC. After
switching to 5-minute cleaning, pressure drop was initially 7 inches WC, but by the end of
the 8-hour test, pressure drop was over 10 inches WC. With conditioning and normal dust
loading {Run 400), the initial bag-clem_g interval was set at 20 minutes, but it was reduced
to a 10-minute interval after 2 hours when the pressure drop exceeded 8 inches WC.
" However, pressure drop =_-aduailyclimbed again and reached 9 inches WC by 5.5 hours into
the test. The cleaning interval was then set at 5 minutes, which resulted in a pressure drop
' of 6 inches WC for the remaining 2.5 hours of the 8-hour test (see Figure 55). Comparing
the last 2.5 hours of each test shows a clear reduction in Ap with conditioning when the
cleaning frequency was 5 minutes for both tests. The K_.values, shown in Table 8, are about
two times lower with conditioning and are reasonably close to the K, values from the tests at
4 and 6 ft/mim
16 Big Brown Texas Lignite

A/C Ratio ffi 16 ft/min .....
14 I ...... e_etme i ]
_.__:c._=_=.____--=---2_n_--_-_-:_._-
s-°--,-_
-]
...........
: ""- :_.... ; ...... _,;...:.,v
.,..... ",,"r'J
_" 10 : ",,::.. t. :, /. I ,'.::'"" I
..........
:
A ,----=,..
: "'_...
-,_-. ........... '_e_,a--e_.,,=.;_
-, ........ .-.---;
$- .,... : !.. ": . _y_l, ,
.. / \A
6
'- 4
e,e
•, I
,;,m&_._ .. ", .. "
.................. .
.a.*.t ..... .... ....
V"--'" .... _ v-"__ ".'_7"v7_ 'i_,,/_- _o -,_
". ..:
• IOlI$ i
k , I ' i w I I
0 100 200 300 400 500

Time (rain)
• Figure 55. Effect of conditioning on baghouse pressure drop for 8-hour Big Brown tests
with a full dust loading at an A/C ratio of 16 ft/rain for Runs 399 and 400. Bags
were pulsed every 5 minutes during the last 2 hours of each test.
73
Inspection of the bags after each test revealed almost no residual dust cake on the
fabric surface for the baseline te_, but there was some dust on the tube sheet, on the clean
side of the bags, and on the cages, which indicated dust penetration into and through the bag.
With conditioning, there was a heavier residual dust cake that was nodular and discontinuous.
However, for the conditioning test, the residual dust cake appeared to be mainiy on the
surface of the fabric, while, for the baseline test, the dust cake aFpearedto be packed into
the fabric structure. One factor that may have contributed to the heavier residual dust cake
with conditioning is the effect of taking the baghouse off-line for about 20 seconds each time
the bag was pulsed. The ammonia and SCh were injected into the ductjust upstream of the
baghouse at a location that is downstream of the inlet isolation valve. Since NH3and SO3
were injected continuously, including the 20-second off-line periods, there was a 20-second
period during each cleaning cycle when ammonia and SO3were injc_axi with no flue gas ..
flow in the duct. Therefore, when flue gas was again introducedto the bagbouse after bag
cleaning, there was an initial surge of higher-concentrationcondkioningagents. The ash that
was exposed to the higher concentrations of conditioning agents could have been more sticky
e
and adhered to the bag, causing a heavier dust cake weight. He attemptwas made to
increase the pulse pressure or durationto achieve better bag cleaning. However, at 16 f't/min
with conditioning, higher pulsing energy may be necessary to adequately control the residual
dust cake weight.
Average particulate emissions were about nine times lower with comlitiomg, achieving
99.97% removal efficiency compared to only 99.79_ for the baseline test. However, by the
end of the 8-hour tests, respirable mass emissions were about two ordersof maguitude lower
with conditioning, as shown in Figures 56 through 58. Therefore, conditioningresulted in a
clear advantage for these full dug-loading tests at 16 ft/rain.
For the two reduced dust-loading tests, particulate emissions were as low as emissions
with conditioning for the normal dug-loading test, but because of the much lower inlet dust
loadings, the calculated colle_on efficiency was only 99.0 to 99.4 %. Respirable mass
emissions for Runs 401 and 402 are shown in Figures 59 and 60. A major bag blinding
problem occurred for both the baseline (Run 401) and conditioning (Run402) reduced dust-
loading tests. For the baseline test, the bag was pulsed after an initial filtration period of
3 hours when the pressure drop reached 6.5 inches WC. The bag cleaned down to 3 inches
WC, but after only 20 minutes, pressure drop reached 7.5 inches WC, and the bag was
pulsed again. This time the bag cleaned down to 5 inches WC and reached 8.5 inches WC
after 15 minutes. The third time the bag was pulsed, the bag cleaned down to 7 inches WC,
so the bag was taken off-line and pulsed 10 times. This brought the pressure drop down to
74
• 102 i .........
101 A/C ffi 16 ft/rain
100 v_ ......
"_............
.o. o
• r_ 10-1
o
............ _ ...............
u .....................
._ ......................................................
104 ...............................
__ i 0 Baseline ]
._ i • 24ppmNH3&12ppmSO3 J •
o, 10-3 ....:..........................................................................................................
104 , . , _ . _ ,,., s . j ',

0 lo0 200 3O0 400 500
Time (min)
Figure 56. Effect of conditioning on average respirable mass emissions for 8-hour Big
Brown tests with a full dust loading at an A/C ratio of 16 ft/rain for Runs 399
and 400.
¢q _., 103 'Big Brown Texas Lignite '

A/C 16 ft/rain
102 0 -0_ _ ..... '0 ...........
_ .......... ........... 0 ____.._.___.,,..--...-_-..-.,_..-

0 O0 O0
101 O°Oo-_'"_O @ 000 , o° C,_'c__
i_ 10° _.o.--co_-o5 5 0"'%--°6v .......
10"
lO-I= ,.._._...v.. m ....................................................................
e. 10"3 .........
• ......................................................
] O After Bag Cleaning i..........
"_" I • Prior to Bag Cleaningl
_. 10.4 ...................................................................
-5
I0 ' ' I ' _ ".... ,' ' I ' '
0 100 200 300 400 500
Time (rain)
, Figure 57. Respirable mass emissions before and after bag cleaning for the 8-hour baseline
Big Brown test with a full dust loading at an A/C ratio of 16 ft/rain for Run 399.
75
• 103 _[ Big BrOwnTexas Lignite ......
" _" i___c = 16_
xo'_........ f o .,,f,,,,s,s
c_,i,,_ I I
t ..._ ,., / , Prio,.
msagcl,,,_,,sI I
,',o,
1oo
104
10"s
0 100 2OO 3OO 4OO 500
Tune (n_)
Figure 58. Respirable mass emissions before and after bag cleaning for the 8-hour Big
Brown test with conditioning and with a full dust loading at an A/C ratio of
16 f't/minfor Run 400.
102
_" Big Brown Texas Lignite
101 A/C = 16 _min

_Baseline .................... _ ................
10o O. ._._0 .0
.-_
e_
i_1 ' _0
- 10" O -
m 0 0 AfterBagCleaning
10.2 -.. ............................ _ _• ;.......
Before Bag Cluning
_,
-_ : _ • • .,, • .O
10"4"' • ' , , , , ", , , ' , "" , ....'"u ' '

100 150 200 250 300 350 ,400 450 500
Time (min)
Figure 59. Respirable mass emissions before and after bag cleaning for the 8-hour baseline
Big Brown test with a reduced dust loading at an A/C ratio of 16-ft/min for e
Run 401.
76
10: _ ..... Big Brown Texas Lignite'
6 ppm NH_ & 3 ppm SOs
101 A/C = 16T't/min
. _
o 0 0 0 0
=o 10° o-o .............................
•_
rd_
o'---'-- o
osm,
' E 10q ...............

o ..........................
.o..... ....................................................................
o
rm 0 AfterBag Cleaning
cii
• Before Bag Cleaning
.,:...............
10.2 ................................................................................................
• • •
• 10-_ ...... .........
.......•
.e. e
......
10 4 ,,, , , , , , ,,i, , ,, ,
100 150 200 250 300
Time (min)
Figure 60. Respirable mass emissions before and a_r bag cleaning for the 8-hour Big
Brown test with conditioningand a reduced dust loading at an A/C ratio of
16 ft/rain for Run 402.
3.5 inches, but it quickly climbed to 7.5 inches WC after 1Ominutes of filtration. At this
point, the pulse-air reservoir pressure was increased to 80 psig (from the 40 psig used
previously), but the bag still would not clean with single off-line pulses. Finally, the bag
was pulsed ten _:ne,s off-line every 10 minutes, which resulted in the bag cleaning down to
about 4 inches, but pressure drop climbed to over 10-inches WC after 10 minutes of f'fltration
time. There was almost no dust cake visible on the surface of the fabric, so an explanation
for the high pres_;uredrop is that ash must have packed very tightly into the fabric structure,
causing irreversible fabric blinding. For the reduced dust-loading test with conditioning
(Run 402), results were very similar in that once high-pressure drop occurred, the bag could
not be adequately cleaned even with increased pulse pressure and multiple off-Linepulses.
• Boththe_tialandeffective
K: values
wereveryhigh,whichisinagreement
withthe
pressuredropdata.The K: valueforthe3-hour
initial
filtration
poriod
forthebaselinetest
, was 35 inches
ofwater-ft-min/Ib,
andtheK,.valuefortheconditioned
testwas 48 inches
of
water-ft-min/Ib.
Thereissome uncemintyoftheexactdustloading
duringthetimeperiod
77
li ' " I, , ,r , '' Iii' ii i .... rlirl , " "ii ' ' ,,' ........ : i,' ii .... I, ii ' ' , I
for which the K2values were calculated, so no difference between them should be inferred.
However, these Ks values are much higherthan the initial Ks values for ali of the full dust-
loading tests, including those conducted at 16 ft/rain (see Table 8). By the end of the low
dust-loading tests, the effective Ks value was 255 inches of water-ft-min/lbfor both the
baseLineand conditioned tests. These dataindicate that pressure dropproblems occurred
from the startof the test and then grew progressively worse after pulsing was initiated.
From the Carman-KozenyKs model, a higher Ks value must be caused either by a smaller
particle size or a low porosity ors combination of both. First, consider the particle-size
data.
Particle-siz_measurements of the fly ash included Coulter counteranalysis of both the

residual dust cake and baghouse hopper ash, and multicyclone sampling of the inlet fly ash.
Coulter counter analysis of ash taken from the bags showed that the volumetricmed_
diameter (VMD) was %0/_m, which _s in the range of 6 to 10/_m for the VMI) of residual
dust cake from the other tests. Therefore, a smaller particle size of the residual dust cake is
apparentlynot the explanation for bag blinding. The indicated VMD of the ash taken from
the baghouse hopper for the low dust-loadingtests was about 26/_m, which is larger than the
range of I 1 to 19/_m for the VMD measuredfor the other tests. The 26-/_m VMD
measurement may not be valid because only a small amount of dust was collected in the
hopper (150 to 200 grams for each test). The possibility exists that the sample could have
been contaminated with residue from the baghousehopper. Another contributingcause for
the large VMI) of the hopper ash is that the loss on ignition (I.O1)for these samples was
3.5 % to 4 % compared to 1% or less for the other te_ (see Table 9). Higher LOI implies
that unburned char was present in the ash, which may have biased the particle-size
distribution toward larger sizes. Multicyclone analysis of the inlet fly ash indicatedthat the
particle-size distribution consisted of a larger percentageof submicronparticlesthan the
previous full dust-loading tests with Big Brown coal, but this difference alone does not
explain the blinding (graphsof multicyclone data are presented in Section 5.4.5.1).
Therefore, a significantly smaller particle-size distribution does not appearto be the primary
cause for the bag blinding and high K,, values.
If particle size is not the explanation, then low porosity or dense packing of the dust
cake must have occurred. Greater dust cake packing is expected at a higher velocity;
however, since the Ks values for the full dust=loading test at 16 ft/rain were not unreasonably
high, the cause for the high K: cannot be attributed only to the 16-ft/min face velocity.
Rather, it appears that the combination of a very low dust loading and high velocity may be
the cause. The particles that are collected are not protected by the rapid buildup of a dust
78
cake and may be more susceptible to movement and packing after initial coUection, leading
tohighKs values.
There is also a question of whether the effectiveness of the bag cleaning is aggravated
, by the combination of high velocity and low dust loading. We know that some dust was
removed from the bags during pulsing because of lowered pressure drop. The amount of
• dust on the bags (based on bag weights before and after the run) and the amount in the
hopper were each about 50% of the total dust collected. "rh_e possibilities exist for the dust
that was removed with pulsing: l) it could be recollected on the bags, 2) it could collect on
the wails of the baghouse vessel and hopper, or 3) it could fall to the hopperbottom. In the
first case, even though pulsing was off-Line,the 23-second off-line time would not be
adequate for individual particles to fall to the hopper. For example, the gravitationalsettling
velocity of a 10-_m particle is only on the order of 1 cm/s. Therefore, ff the dust is
dispersed as individual particles after pulsing, these particles would not be expected to reach
the hopper. For dust to reach the hopper after pulsing, sufficiently large agglomerates must
remain intact for the dust to overcome the viscous drag of the flue gas. If the combination
of a high velocity and low dust loadin(_causes the particles to be more completely disbursed
after pulsing, poor cleaning and event_ bag blinding may occur.
An additional
factor
that
mightcontribute
todustcakepacking
andbagblinding
isthe
cohesive
nature
oftheash.Ifcofiring
ofthecoalwithnatural
gascaused
theflyashtobe
significantly
lesscohesive,
theresults
wouldbe moreeasily
explained.
Unfortunately,
not
enoughashwas generatedtoconduct
tensile
strength
orporosity
measurements
on thebulk
flyash,sono dataon cohesive
properties
aleavailable.
One additional run was conducted at 16 ft/rain with Big Brown coal 1_odetermine ff
precoating the bag with Neutralite® might prevent bag blinding. Again, the combustor was
fired on natural gas, supplemented with a coal feed rate of approximately 1/20th of the
normal rate. This resulted in a dust loading at the baghouse inlet of 0.22 grains/scf. While
the Neutralite® provided some initial benefit, after the bag was pulsed several times, it
appeared that most of the Neutralite® was removed, and bag blinding again occurred. The
bag blinding occurred for both the baseline and conditioning tests.
• The bag was initially precoated with Neutralite® while firing the combustor on 100%
natural gas up to a AtP of 2.5 inches WC. The test was started with a low A/C ratio to help
' establish a dust cake without fabric blirlding. Filtration on coal was started with an A/C
ratio of 4 ft/min for one hour followed by one hour at 6 ft/rain. At this point, the AtP was
79
6 inches WC, primarily because of the Neutmlite ° precoat, so the bag was pulsed off-line.
Visual inspection of the bag through the sight ports indicated that almost all of the dust,
including the Neutralite® precoat, was removed by the pulse. The A/C ratio was then
increased to 8, 12, and finally to 16 f't/min. At this point, the decision was made to pulse
the bag at 5 inches WC instead of on a time basis. After the first pulse at 16 ft/rain, an
entire hour passed before reaching 5 inches WC, which indicated some continued benefit of
the Neutrafite ®. The time required to reach 5 inches WC for the next cleaning interval was
only 30 minutes; the time required for _e, foUowing interval was 20 minutes; and the time
for the next cycle was reduced down to 12 minutes. This indicated that the benefit of
Neutralite ° was being lost with each successive cleaning cycle, and the bag was again
blinding. After reaching 5 inches WC within 10 minutes, the bag was considered to be
blinded because this is equivalent to an effective K2 of about 75 inches of water-ft-rain/lb.
At this point, the bag was removed, hand-cleaned, and again precoated with Neutralitea.
However, after about four or five cleaning cycles, the benefit of the Nentralite ° was quickly
lost, and the bag had to be cleaned at least every I0 minutes to keep the _P below 5 inches
WC. Precoating with Neutralite* after hand-c!eaning was tried several times, but each time
' blinding occurred after only four or five cleaning cycles. Tests were also conducted in
which some Neutralite° was added after each pulse, but this did not prevent bag blinding
either. In cases where Neutndite a was added simultaneously with the fly ash, the _P did not
increase as fast; however, the add rate of the Neut_ had to be greater than the fly ash
dust loading to prevent rapid _P increase. Whenever the Neutxafite ® was stopped, the
benefit was quickly lost, apparently because most of the Neutmlite ° was removed with bag
pulsing.
A test was also conducted with 613 ppm of HH,JSO3 with a Neuualitea precoat, but no
benefit of the conditioning was observed. A final test was attem_ in which a Neutralite0
precoat was used, but the coal feed rate was increased to the 16-1blhr from the 4-1b/ht rate
used in the earlier tests. However, fabric blinding again occurred after a few cleaning
cycles. Since the Neutralite® was not successful in preventing fabric blinding, the run was
terminated, and the remainder of the week was used to perform another test with the Black
Thunder coal (Run 413).
The reasons for the bag blinding with a low dust loading (achieved by cofiring with
natural gas) remain somewhat of a mystery. Limited particle=size data do not indicate that
the problem was caused by a smaller particle-size distribution. A sufficient amount of ash
was not available from the f'trst low dust-loading tests (Runs 401 and 402) to conduct
cohesive measurements, and the ash from Run 412 was highly contaminated with Neutralite ®,
80
so cohesive dataare not available. Whether the cohesive propertiesof the dust from the low
dust-loading tests were different than ash properties from the full dust-loadingtests is not
known. The remaining explanation is that, under very low dust-loading and high-velocity
conditions, the ash is apparentlymore subjectto packing, which leads to fabric blinding.
h
5.4.5 C0he_jveProperties and Ash Analyses

q
5.4.5.1 Particle Size
Particle size was measured by two main methods: Coulter counter and 6-stage
multicyclone. Typically, one Coulter counter analysis was completed on a hopper ash
sample from each run. The Coulter counter data arc presented as volumetric median
diameters (VMD) in Table 9. The VMD of the hopper ash samples typically ranged from 12
to 16/zm, with one outlying at 19 pm. The Coul_r counter analyses arc fairly accurate, but
caution must be exercised in interpretingdifferences in the VMD among the different runs.
Since Coulter counteranalysis requires only a very small sample (on the order of
milligrams), a question exists on how representativethe small sample is of the entire hopper
ash sample. Multiple analyses would provide more informationon the variability of the
VMD within a given run, but multiple analyses were not completed. Therefore, no
differences should be inferred in the VMD values for the hopper ash samples. The MI:)
values of the residual dust cake samples, however, were clearly smallerthan the VMDs of
the hopper ash samples and typically ranged from 6 to 8 _,m, with one value at 10 _m. This
indicates size fractionationoccurrezibetween the hopper ash and the residual dust cake ash,
which could be the result of several effects. Largerparticles might be more easily removed
from the bags because the ratio of drag or inertial forces to cohesive forces is likely to be
greater for larger particles. In addition, for the particles that are removed with pulsing, the
larger particles are more likely to reach the hopper rather than be recoUectedcn the bags.
Both effects would tend to make the residual dust cake finer than the bulk hopper ash.
The multicyclone analyses are shown in Figures 61 through 64 as log-probability plots.

A log-normal distribution curve was fit to the data, which is a straight line on a log-
probability graph. The Big Brown and Milner multicyclone data from Task 3 a_ee closely
with the Task 2 multicyclone data (Figures 61 and 62). A comparison of multicyclone
• results for the three coals, shown in Figure 63, indicates some differences in particle-size
distribution. The Black Thunder ash has a greater percentage of submicron particles than the
, other ashes. Extrapolation of the data to the 50% cumulative mass point should provide the
mass median aerodynamic diameter. For the Big Brown and Milner ashes, this aerodynamic
81
Figure 62. Particle-size distribution of Miiner fly ash based on multicyclone sampling.
82
'. 90 i i i ! ! : _ ' : : : " : i
iO BlackThunderTask3,Tests
70 i!3 Big Brown Tas_3ffestsi i
:...........................................................
'" .............
Milner Task 3 Tests i i i i
' _
_,_ 30 i: i i. i ! :(
om ; '. ', " i .....
!!ii: ! .:
1 : : i : i i :. i i
' . ! i i : i
0.I .............
:......
:.....
:.,,.i...i.,.:
........................
i
0.01
1 I0
AerodynamicDiameter(_m)
Figure 63. Comparison of particle-size distributions of Big Brown, Milner, and Black
Thunder fly ash based on multicyclone sampling.
0 i ! ill i i i :
A Big Brown Raduced

70 ,,i..,D
......
.,B,!g...B..r.o..w...n.
.....
.N.,9,_a,!..,D,.C.,_t.
iL.0.a.d!n.g.
50 ...i......i..+._
...... _.,
!ili! !i_i :
! ili! _
_ _ _':_i 'ii
O__" 10 ': •:'": : :
'q_ " ! ! i ! : : : : i
_ i ii _ i :ii:
--- : ! i _i ' i i ! :. i
..... ! _ _
0.I ................. _ i,
. ; : .
0.01
1 10
" Aerodynamic Diameter (_m)
• Figure 64. Comparison of particle-size distributions for normal and r_duced dust-loading
tests for Big Brown, based on multicyclon¢ sampling.
83
median diameter is about 20 _m, or about 13-_m volumetric diameter, which agrees with the
Coulter counter data (aerodynamicdiameter can be convened to volumetric diameter by
dividing by the square root of the particle density). The aerodynamicmedian diameterfor
the Black Thunder ash correspondsto a volumetric median diameter of about 17 _m, which
is slightly higher than the 12- to 15-_m range in volumetric median diameter determinedby
Coulter counter• Extrapolationof multicyclone data, however, is risky because about 80%
of the ash is typically collected in the first cyclone, and no furtherinformationis availableon
the particle size of that 80% other than its being larger than the cut point of the first cyclone.
Both the mass median diameter and the width of the distribution(geometric standard
deviation) are likely to have an effect on K2. The highest K2 occurred with the Black
Thunder ash, which also had the greatest percentage of fine pe_cles, and the lowest K2
occurred with the Milner ash, which had the smallest percentage of fmc particles.
Therefore, even though the mediandiameters of the three ashes are similar, the differences
in the amount of free particles may partially account for differences in K2. Multicyclone
results for the reduceddust-loading tests, shown in Figure 64, indicate a greaterpercentage
of fine particles than the normaldust-loading tests. However, this difference alone does not
appear to be the entire explanationfor the mu'h higher K2 values for the reduced dust-
loading tests.
5.4.5.2 Tensile Stren_h
The primary dust characteristicsthat are likely to affect fabric filter performance
include particle-size distribution and cohesive characteristics. The cohesive chantcteristics
depend on factors such as chemical composition and particle morphology. The cohesive
nature of the dust refers to the "stickiness" of the dust, or, more descriptively, the particle-
to-particle binding forces that exist in a dust cake. It is these relatively weak particle-to-
particle forces that determine the pore-bridging ability of the dust and the resistance of the
dust to ree,ntrainment or gradual seepage through the fabric. These effects are important
because they ultimately determine the fabric filter particulate collection efficiency. The
resistance of the dust to packing and the ease of removing a dust cake from the fabric also
depend on the particle-to particle forces and are important because they determine the
pressure drop. The exact particle-to-particleforces depend on particle size, contact area, and
chemical composition. Subsequently, they are likely to vary widely within a given dust cake.
Measurement of the distribution of individual particle-to-particle forces would be extremely
difficult. A more strmghtforward approach would be to determine the average tensile
strength of the dust, which can be considered to be a measure of the average particle-to-
84
particle binding forces. For dusts with the same particle-size distribution, a comparison of
tensile strengths at constant porosity should be an indication of their relative particle-to-
particle
bindingforces.
Tensilestrength
asa function
ofporosity
forthe 100-hour
Big Brown,Milner,and
BlackThundertests
isshown inFigures65 through68. Basedon multiple
curvefitanalysis,
• an exponential
curvewas fit
toeachdatasetoftheform:
tensile
strength
= a exp ([-b]
[% porosity]) [Eq.3]
The physical significance of the coefficients in Equation 3 is not known. The curve fits
simply facilitate extrapolation of the dam so that comparisons in tensile strength can be made
at constant porosity.
For the Big Brown tests (Figure 65), there is a significant shift in the tensile strength
• with conditioning toward higher porosity and higher tensile strength. For the Milner tests
(Figure 66), there is also a significant shift toward higher porosity and higher tensile strength
with conditioning, but the highest tensile su_ngths observed (under maximum compaction)
were only about 4 gJcm _ compared to 18 g_/crn_ for the conditioned Big Brown samples.
For both Big Brown and Milner conditioning tests, the tensile su_ngt_ at a face velocity of
6 ft/min were somewhat higher than the tensile strengths at 4 f-t/rain. In the B-hour tests
completed in Task 2, the highest tensile strengths were also observed at the higher A/C
ratios. Typically, when the A/C ratio is increase.A, the bags are cleaned more often, and the
exposure time of the ash to the flue gas is reduced. However, the bag-cleaning interval was
held constant for both A/C ratios, so differences in tensile strength would appear to be
related to differences in face velocity, rather than other factors such as exposure time. There
is no apparent reason why conditioning would have a greater effect on tensile strength at the
higher A/C ratio. The shift in the tensile strength curves with increasing A/C ratio was
much smaller than the significant shift in the tensile strength curves with conditioning for
both the Big Brown and Milner tests. This indicates that the cohesive character of the ash
was significantly changed and implies that the improvement in filtration performance was a
direct result of the changed cohesive properties.
" Figure 67 shows that the shift in the tensile strength curves for the Black Thunder tests
was smaller than the shift for the Big Brown and Milner tests. For the Black Thunder tests,
• there was also a smaller improvement in fabric filter performance, which implies that the
level of performance improvement correlates with the relative shift in the tensile strength
85
• 2O i
t_ • Conditioned., lO0-hr, 4 ft/rain

""- • Conditioned, 100.br, 6 fthnin
= 10 i
- i
e_
0
30 40 50 60 70
Porosity (%)
Figure 65. Effect of conditioningon fly ash tensile strengthfor 100-hour Big Brown tests.
. i i i i
! Milner Coal
........ + o _.,.._.,m.,.+
,,+,.,.,,,+
+ i i i
g i i • i
.-: 2 ...............
............................
...................................
+,.+'+
+ + +++
II I " "......... "*" .... **_* "*'"_ _--'*'*'"" "*" _*°'_ "'" ........ "**--_ "" *
o 50 S5
i' i i"-i" 60 6S 70 75
Porosity (%)
Figure 66. Effect of conditioning on fly ash tensile strength for 100-hour Mi.lner tests.
86
" 8 Black: Thunder Subbim'mi_ous Coal i'
7 ....... ..........i ............i
_2
3-
!!!!!!!!!!!!!!!!!
1 ................................
- .....
a° :i---*
....................
::-
..............................
, ,i [ n , i
30 40 50 60 70
Porosity (%)
Figure 67. Effect of conditioning on fly ash tensile strength for 10G-hour Black Thunder
tests.
BT - Black Thunder
. Milner c - Conditioned 4 - A/C-4 ft/rain
- Big Brown bl - Baseline 6 - A/C =6 ft/rain
100 -
i
i t
BT ......
_. 1 _-------------"------_ - "" - MN-c..4""
"_ ' MN-bl-4
BB-bl-4
0.1 -------,------------
30 40 $0 60 70
Porosity, %
. Figure 68. Comparison of fly ash tensile strengths for conditioned tests with Big Brown,
Milner, and Black Thunder coals.
87
curves. An obvious question is whether fabric filter performance can be predicted on the
basis of tensile strength anger porosity measurements. A comparison of the tensile strengths
for the three baseline coals, given in Figure 69, indicates that the three ashes are distinctly
different. Recall that the highest particulate emissions occurred for the baseline Big Brown
run and the lowest particulate emissions occurred with the Black Thunder coal (see Figures
29, 37, and 46). The lowest particulate emissions appear to correlate with the highest tensile
strength, which indicates that particulate emissions might be predictedfrom tensile strength.
A question exists as to what porosity value or compaction force the tensile strength should be
compared, but it appears that either the maximum or average tensile strength for each sample
would provide a relative indication of particulate emissions. This comelation demonstrates
the potential to predict filtration performance basea on ash properties.
Tensile strengthmeasurements, including the 8-hour tests in Task 2 and the 16-ft/min
tests, are shown in Figures 70 through72. Figure 70 shows that the tensile strengthfor the
conditioned ash from the normal dust-loading 16-ft/rain test was within the range of tensile
strengths from the other 8-hour tests. This might indicate that tensile strengthis not
influenced by A/C ratio, as suggestedpreviously; however, since the bag was cleaned much
more frequently, a direct comparisoncannot be made because of the possibility of residence
time effects. For two of the Black Thundertests, tensile strengthof the residual dust cake
ash was measured (Figure 72) to help determinewhy bag-cleaningproblems occurred.
Results show that tensile strengthof the residual dust on the bags was considerably higher
than the tensile strengthof the correspondinghopper ash samples. The highest tensile
strength was measured for the bag ash from Run 410B, where the conditioning
concenuations were 12/6 ppm. This is surprising because the bag-cleaningproblem occurred
when concentrations were 24/12 ppm in Run 410A. However, the residual dust cake was
exposed to the flue gas for only 24 hours in Ran 410A, comlmredto 75 hours in Run 410B.
In addition, the median particle size of the residual dust cake was somewhat smaller for
Run 41OB. The higher tensile strengthsof the dust cake ash may be the result of both a
smaller particle size and a longer time of exposure to the flue gas. The tensile strength of
the hopper ash sample for Run 410A at 6 ft/min, when bag cleaning occurredand
conditioning concentrations were 24/12 ppm, was higher than the tensile strengthfor
Run 413 at 6 ft/min, when bag cleaning did not occur and conditioning concentrations were
12/6 ppm. The higher tensile strengthat the high conditioning concentrationswas expected.
However, the maximum tensile strength for Run 410A, when bag-cleaning problems
occurred, was less than the maximum tensile strength of ash from the Big Brown tests when
there were no bag-cleaning problems. Tiffs implies that other factors might also influence
bag cleanability, but we do not know which dust cake had the highest actual tensile strength.
88
• '° $ i iii i
• '_' ['! I
ii Bi_ Brown B_eline
ii A Black Thunder"Baseline
4 i \--_ i|_ MiIne?-Basehne
°
__ 3 ....................
i\ i
_.....................
__-"'"'""'""_.......................
":
[
........................
"................ ".................
• _ ! _,, ! i i
• _ 2 ....................
i...... i........................
i.................
i.................
•__ 1 ..........
_ !
_'i.................
N
_.......
'_...........
' Ooi
7 i i _ i i | li i
30 35 40 45 50 55 60
•Porosity (%)
Figure 69. Comparison of fly ash tensile strengths for baseline tests with Big Brown,
Milner, and Black Thunder coals.
0 i
• 30 40 50 60 70
Porosity (%)
Figure 70. Fly ash tensile strenThs for 100- and 8-hour Big Brown tests.
89
• 6 i
i
5
A _ondition_l, lO0-1u',6 _min
E o ,Baseline, 8-ht,!4 R/rain
4, 8
Cl Conditioned, 8-111",
"_ 4 "_
,- o
iO .
-- 2
! [] i,
! 0
0
$0 55 60 65 70 75
Porosity (%)
Figure 71. Fly ash tensile strengthsfor 100- and 8-hour bfilncr tests.
12 i
• 24/12 6 f(/m;_._
10
S ,, 12/6_mi 6 _m.
ta II 24/12 t, bag sample
eJo 8
M
= 6
t.,
r._ ,t,
--
e,mm 4
lm
2 m'
®
Oq •
0
30 40 50 60 70
Porosity (%)
Figure 72. Comparison of tensile strengths of hopperash and residual dust cake ash from .
the bags for Black Thunder tests.
9O
One of the difficulties in trying to correlate tensile strength with bag cleanability is the
determination of the actual tensile strength of the dust cake. If the Ks of the dust cake is
known and K: can be measured in the laboratory as a function of porosity, the porosity of the
dust cake can be determined ind_tly, If the tensile stren_h of the fly ash is measured as a
. function of porosity, the tensile strength of the dust cake can be inferred. This approach
assumes that the hopper ash and dust cake have similar particle-size distr,_butions.
. Laboratory measurements of K_ as a function of porosity arc shown in Figures 73 through
76. These combined data from both baseline and conditioning runs for each coal arc fit to
the Carman-Kozeny equation for each plot. The Big Brown data ali follow a single Carman-
Kozeny curve more closely than the Milner and Black Thunder data, but the dust cake
porosity can be approximated from these curves if the K: of the dust cake is known. One of
the causes for data scatter is small differences in particle-size distribution. Figure 76 shows
the Carman-Kozeny curves fit to the three sets of data along with calculated Carman-Kozeny
curves for monosized panicles. Note that the differences in the curve fits are equivalent to
only a 2-_m spread in equivalent monosized diameter. Since the actual dusts have broad
particle-size distributions, the equivalent Carman-Kozeny diameters cannot be predicted
direcdy from the median particle diameters of the dusts. From the initial K2 values for the
100-hour tests (corrected for the viscosity difference between 70* and 300*F) and the K2
porosity curves in Figure 76, the porosities of the baseline and conditioned Big Brown dusts
cakes were 63 % and 77 %; the porosities of the baseline and conditioned Milner dust cakes
were 80% and 85 %; and the porosities of the baseline and conditioned Black Thunder dust
cakes were 66 % and 72 %. These porosities are ali much greater than the porosity range for
which tensile strength data are available. Therefore, to infer the tensile strength of the dust
cakes would require an extrapolation greater than the porosity range for which the tensile
strength data exist. Because of the data scatter of the actual tensile strength data, such
extrapolation would have too much uncertainty to provide valid tensile strengths of the dust
cakes. From the available data, we cannot accurately infer the tensile strengths of the dust
cake. One conclusion that can be made, however, is that the actual dust cake tensile strength
would be less than the lowest measured tensile strength for each ash sample.
5.4.5.3 Porosity Measurements
)
The aerated and packed porosity measurements are most likely related to the tensile
strength, but may provide more direct information on the susceptibility of the dust to
)
packing. Aerated and packed porosity data, given in Table 8, clearly show that one effect of
conditioning is to make the ash less susceptible to packing. The aerated and packed porosity
data are in good agreement as to the increase in porosity that occurs with conditioning. For
91
30 '
25 .c3
-_
m 0 Baseline
== O @ 24ppmNH 3&12ppmSO 3
20
Y O
!
o
o
o
__
10
@mm
eq
0
20 40 60 80 I00
Porosity (%)
Figure 73. Laboratory measurement of K_ as a function of porosity with Carman-Kozeny

equation fit to data for Big Brown tests.
25 Milner Coai o h.e_u, '

I • 24 ppm NI_ & 12ppm SO 3
-_ 20
m
mm
-.
15 "--
%
mm
_J l0
"=
0 " ' I ' i ' ' 1 ' "
20 40 60 80 100
Porosity (%)
Figure 74. Laboratory measurement of K2 as a function of porosity with Cannan-Kozeny
equation fit to data for Milner tests.
92
40
Black Thunder Subbituminous Coal
35
m
30
e_ 0 Baseline
• 12 ppm NI_
............... & 6 ppm
_.o_oo..oSO3
_. 25 z_ 24 ppm & 12 ppm S0 3
'_ 20
. .= 15
em
0
20 40 60 80 100
Porosity (%)
Figure 75. Laboratory measurement of K2 as a function of porosity with Carman-Kozeny

• equation fit to data for Black Thunder tests.
- .. . . ...
4035t _i \\\ \'\ \\ D_.\ D-3,=. Carman-KozenAsparticlea
FunctiOnsize_-_of
"
E 25 o.o. ! \\ \
\ ,., \\_ \_._ . . ..... Big Brown
"4 ',. ; •
--¢_
m 20 : ,_ \_,...._.'..
_, ., " . ----- Milaer
--....- Black Thunder
; '_ "• •
._ 15 . ._ ..
5
;i::::::....
0
20 40 60 80 100
Porosity (%)
Figure 76. Comparison of K, curves for Big Brown, Milner, and Black Thunder tests.
93
example, the greatest increase in aerated and packed porosity that occurredwith conditioning
was with Big Brown coal, and these tests also resultedin the greatest decrease in K2. The
smallest increase in aerated and packed porosity as a result of conditioning was with the
Black Thunder tests, which also resulted in the smallest decrease in K2. The correlations
between the aerated and packed porosity data and the initial K 2 ValUeSfrom the 100-hour
tests are shown in Figure 77, and the correlationsfor the effective K2values are shown in
Figure 78. It atvears that the packed porosity data provide a somewhatbetter correlation
than the aerated porosity data. These correlations would likely not be valid, however, if the
particle-size distributionsvaried significantly. The actual porosity of the dustcake is most
closely represented by the aeratedporosity measurements,rather than.the packedporosity
measurements. For example, with the Big Brown tests, the inferredporosities of the baseline
and conditioned dust cakes were 63 % and 77 %, comparedto the aeratedporosities of 58 %
and 76 %. For the Milner tests, the inferreddust cake porosities were 80% and 87%,
compared to aeratedporosities of 71% and 83%. For the Black Thunder tests, the inferred
dust cake porosities were 66% and 72 %, comparedto aerated porosities of 66% and 70%.
Therefore, the aerated and packed porosity measurementshave three uses in relating ash
properties to f'fltrationperformance. First, they appearto be good methods to quantify the
effect of conditioning on ash properties; second, they correlate well with initial and effective
K2 values; and third, the aerated porosity measurementsprovide an approximationof the
actual dust cake porosity. Much more work is needed, however, to refine these correlations
and to generalize them for a variety of dusts.
5.4.5.4 AmmoniaandSulfateAnalyses
Elemental analysis of the hopper ash samples, determined by x-ray fluorescence, is

shown in Table 9. The transferof SO3to the fly ash is shown by the in_ sulfate in the
conditioned samples comparedto the baseline samples. However, this analysis is not
sufficiently accurate to determine the amountof the injected SO3that was transferredto the
fly ash and does not include ammoniaanalysis.
Ammonia and soluble sulfate analyses Coyion chromatography)were conducted on the

baghouse hopper ash samples and on the bag ash samples and are presented in Table lO.
Since ammonium sulfate is highly soluble, this shouldbe a valid determinationof the
amount of conditioning agents transferred to the fly ash. The ammonia in the ash and
ammonia slip in the flue gas downstream from the baghouse should equal the amount of
ammonia injected. For the Big Brown tests, the average of ten ammonia slip measurements
was 0.83 ppm and ranged from 0.27 to 1.84 ppm. For the Milner tests, the average of six
94
• 10
o Aerated Porosity
• Packed Porosity
omm
=
T
• 0
osm 2 ..........
ImU
0 ' , i ' I _' I ' I ' " I "
30 40 50 60 70 80 90
Porosity(%)
Figure 77. Correlation between initial K2and aeratedor packedporosity.
0 i i i i|, , i . i i
= o Aerated Porosity
=
ommt
• Packed Porosity
=
_" 15 ................
t ...........................................
o ......
!
%
10 ................................
=
,Diii
eq
•_" o o
• 30 40 50 60 70 80 90
Porosity (%)
Figure 78. Correlation between effective K2 and aerated or packed porosity.
95
96
".o
ammonia slip measurementswas 0.18 ppm and ranged from 0.10 to 0.31 ppm. For the
Black Thundertests, the average ammoniaslip for seven measurementswas noticeably
higher at 4.9 ppm and ranged from 4.3 to 6.7 ppm. This was true in spite of the fact that
five of the seven ammonia slip measurementswere conducted when the ammoniainjection
• rate was only 12 ppm. Average ammonia slip values for each run are shown in Table 10.
• For the Big Brown and Milner samples, most of the injected ammoniaand 503 was
transferredto the fly ash, based on soluble ammonia and soluble sulfate analyses of hopper
ash samples. There is uncertaintyas to the sulfate transferredbecausethere were also some
soluble sulfates _ the baseline ashsamples. The basefine sulfate concentrationswere
subtracted fromthesulfate concenwations withconditioning toinky determine the
amountofaddedsulfate. ForthelO0-hour tests,
theamountoftheinjected S03 thatwas
transferredto the fly ash rangedfrom 79 % to 110%. The uncertaintyof the amount of
injected SO3 of about ± 10% along with the uncertaintyof the amountof baseline soluble
sulfate means that the uncertaintyof the sulfatebalance is at least 4-10%. For the 8-hour
Big Brown test at 16 ft/rain (Run400) and the Black Thundertest at the highest injection rate
(Run 410A), the soluble sulfate concentrationswere much lower than'e_. The
ammonia mass balance was based on the ammoniaslip and the ammoniain the hopper ash
and ranged from 76% to 104% for all tests except Run 400, for which the balance was 64%.
The ammonia balances should be more accuratethan the sulfate balancesbecause there is
greater certainty of the amount of injected ammonia, and there is no backgroundammonia in
the baseline ash. The ammoniaandsulfate concentrationsfor the residualdust cake samples
were from 2 to 5 times greater thanthe correspondinghopper ash samples. However, this
difference is not surprisingconsidering the large difference in particle size. Since the
conditioning agents are likely to deposit on the available surface area, the ammonia and
sulfate are expected to be more concentratedin the finer particles.
The measured ammonia-to-sulfatemolar ratio for the 100-hourMilner tests was close
to 2:1, indicating that the productformed on the ash is ammoniumsulfate. For the 100-hour
Big Brown tests, there was also a molar ratio of ammonia to sulfate of 2:1 for the hopper
ash. However, the fly ash ammoniaconcentration,along with the 1 to 2 ppm of ammonia
slip, still leaves 4 ppm of ammonia unaccountedfor from the total of 24-ppm ammonia that
J
was injected. Whether some off-gassing of the ammonia could occur in sample storage and
handling or whether the lack of complete closure is due to analyticalvariability is not known.
For the Black Thunder samples, the ammonia analysis indicated that only about 30 %to
50 % of the ammonia was transferredto the fly ash, and the ammonia slips were
97
correspondinglyhigh. About2 to 4 ppm of the injected ammonia was unaccountedfor. One
other interesting effect that occuned with the Black Thundersamples was very low soluble
sulfate values, on the basis of a 2_hour leaching test. Valid soluble sulfate concentrations
can be obtained if the sulfate measurementsare taken immediately after water extraction, but
the soluble sulfate concenu'ationslowly decreases with time. Apparently, insoluble sulfate
complexes are slowly formed in solution. The molar ratio of ammonia to sulfate was closer
to 1"1 for the 100-hourBlack Thundertests comparedto a 2:1 ratio for the Big Brown and
lVfilnertests. This implies that there could be a reaction between the fly ash and SO3in the
dust cake and often an explanation for the high ammoniaslip values. The high ammonia
slip, l"1 molar ratio of ammonia to sulfate, and the less significant effect of conditioningon
baghouse performancefor the Black Thunder tem areindications that the effectiveness of
conditioning depends on fly ash chemistry.
In addition to measuringthe ammoniumand sulfate concentration,the pH of the

residual dust cake ash was measuredfor eight of the 100-honr tests and two 8-hour tests (one
gram ash in 100 ml of water). As shown in Table 10, there is a decrease in the pH of the
ash with NH3 and SO_conditioning. As would be expected, the decrease in ash pH is
directly related to the alkali content of the coal ash (Ca, Mg, Na, and K). The Black
Thunder, which had the highest alkafi content (34.4 _) of the three coals tested, showed
almost no decrease in pH. Contr_g this to the kfilner fly ash, with an _ content of
only 7.5 %, the pH dropped from 11 for the baseline test to 6.6 with conditioning. This is
very near the pH of 5.5 for a 5 % solution of pure ammonium mlfate, lm" the Big Brown
coal, which has an intermediatealkali content (27.4_), the pH _ from 11.9 for the
100-hour baseline test to 9.6 for the 100-hour conditioned test. The decreasein pH also
seems to be nm-time-d_t. For the 8-1muftest using Big Brown coal, the pH
decreased only slightly, 12 for the baseline to 11.3 with conditioning. This indicates that
there is an increase with time in the ammoniumsulfate (bisulfate) contentof the residual dust
cake ash. This is consistent with the higher ammoniaand sulfate concentrationof the
residual dust cake ash for the 100-hour tests. What, if any, effects this may have on long-
term operation is uncertain.
5.4.5.5 Residual Dust Cake Weight
Residual dust cake weights, given in Table 8, were determined by measuring the bag
weights before and after the runs. After each 100-hour run, the whole tube sheet was lifted
with the bags in piace to allow for inspection of the bags and photographs of the residual
dust cakes. The cages were then pulled from the bags, and the bags were removed from the
98
botton, of the tube sheet to minimize the amount of dust loss prior to weighing each bag.
Photographs of the bags are shown in Figures 79 through 87. The heaviest residual dust
cakes occurred for the Big Brown tests, and the tightest cakes for the Milner tests. For the
baseline Big Brown test (Run 403, shown in Figure 79), there was very little dust on the
fabric surface over the entire length of the bags, but the dust was somewhat packed into the
fabric structure. For Run 404 with conditioning at an A/C ratio of 4 ft/rain, the weight of
the residual cake was slightly lower, but the dust was very much on the fabric surface and
somewhat nodular on the bottom half of the bags (Figure 80). For Run 405 with
conditioning at 6 ft/rain, the cake was heavier and even more nodular on the bottom half of
the bags. The top 8 inches of the bags with conditioning had a very heavy cake about 1/2
inch thick. Some of this heavy cake fell off when the tube sheet was lifted, but it is clearly
visible in the photographs. Apparently, the pulse results in tittle bag expansion or reverse
flow at the very top of the bag, so this dust buildup may have occurred over ali the runs.
For the Milner tests, the photographs show almost no residual dust cake for either the
baseline or conditioning tests, which is in agreement with the bag weight data. In spite of
" the extremely high dust loading with the Milner coal, the ash characteristics apparently
: facilitate bag cleaning. However, a different fabric was used for the Milner (P84/Ryton),
which may also have had an effect on the dust cake release.
I
The bags were also fairly clean at the end of the Black Thunder tests, but there was
some nodular structure on the bottom half of the bags. They were not as clean as for the
Milner tests, but were cleaner than for the Big Brown conditioning tests, which is in
agreement with the residual dust cake weights.
6.0 CONCLUSIONS AND RECOMMENDATIONS
6.1 8-hour Tests with Milner and Big Brown Coals
Results from the 8-hour tests showed that conditioning significantly reduced pressure
drop and corresponding K 2 at ali concentrations tested. Results showed that conditioning
worked well for both the MiMer and Big Brown coals, even though the two coals had very
• different properties. This is an encouraging result and may be the most significant finding of
the Task 2 tests, because it demonstrates the wide applicability of the conditioning process
• for improving baghouse performance.
99
Figure 79. Residual dust cake at the end of the 100-hour baseline test with Big Brown coal
for Run 403.
100
• Figure 80. Residual dust cake at the end of the 100-hour conditioning test with Big Brown
coal at an A/C ratio of 4 ft/min for Run 404.
101
Figure 81. Residual dust cake at the end of the 100-hour conditioning test with Big Brown
coal at an A/C ratio of 6 ft/min for Run 405.
102
• Figure 82. Residual dust cake at the end of the 100-hour baseline test with Milner coal for
Run 406.
103
Figure 83. Residual dust cake at the end of the 100-hour conditioning test with Milner coal
at an A/C ratio of 4 ft/min for Run 407.
104
• Figure 84. Residual dust cake at the end of the 100-hour conditioning test with Milner coal
at an A/C ratio of 6 ft/rain for Run 411.
105
Figure 85. Residual dust cake at the end of the 100-hour baseline test with Black Thunder
coal for Run 409.
106
• Figure 86. Residual dust cake at the end of the 100-hour conditioning test with Black
Thunder coal at an A/C ratio of 4 ft/min for Run 410B.
107
Figure 87. Residual dust cake at the end of the 100-hour conditioning test with Black
Thunder coal at an A/C ratio of 6 ft/rain for Run 413.
108
Results showed that conditioning was more effective for the woven glass fabric than for
the Ryton felted fabric, but it was highly effective for both fabrics. Particulate emissions and
pressure drop were both substantially reduced with the woven glass fabric with Big Brown
coal at an A/C ratio of 4 ft/mm. With the Ryton fabric, pressure drop was substantially
. reduced, but particulate emissions were reduced only for the test at an A/C ratio of 6 ft/min.
However, collection efficiency was already high at 99.99%, and some additional reduction in
• particulate emissions would be likely if the bags were cleaned on a pressure drop basis rather
than a constant time interval. Since pressure drop was reduced by over 50 % for the Ryton
fabric with conditioning, a longer cleaning cycle interval could have resulted in lower
emissions because of fewer bag-cleaning spikes in a given time interval. However, the
Ryton fabric already appears to provide adequate particulate collection efficiency at an A/C
ratio of 4 ft/rain. For the Milner tests, there did not appear to be a significant difference in
the performance improvement seen with conditioning between the Ryton and P84/Ryton
fabrics.
Comparing the two Big Brown tests at an A/C ratio of 6 ft/min provides strong
evidence that conditioning will greatly facilitate operation at high A/C ratios. While no
baseline tests were conducted at an A/C ratio of 8 ft/rain for direct comparison, the highly
successful results from conditioning tests at an A/C ratio of 8 ft/rain for both coals further
imply that this is a major benefit of conditioning. In general, conditioning provided a
reasonable pressure drop and excellent particulate collection efficiency for both the Big
Brown and Milner coals when the pulse-jet baghouse was operated with Ryton fabric up to
an AJC ratio of 8 ft/min. Results imply that operation at even higher AJC ratios should be
possible when using ammonia and SO3 conditioning. The extent to which A/C ratio can be
increased with flue gas conditioning was further evaluated in Task 3 tests at an A/C ratio of
16 ft/min.
The primary effect of conditioning on ash properties appears to be a significant increase

in aerated and packed porosity. Conditioning produces an increase in the tensile strength at
constant porosity, which implies that the particle-to-particlebinding forces are increased.
While some new information was generated during Task 2 testing, the best
concentrations to use remained somewhat uncertain. The Big Brown tests showed that
significant performance improvement was achieved for ali three concentrations tested, which
is an encouraging result because it shows that performance improvement is not highly
• sensitive to exact concentrations of conditioning agents. Perhaps the most significant result
of the conditioning agent concentration tests with Milner coal was the excellent performance
109
' o,
improvement that was achieved with the relatively low concentrations tested, even though
this was an extremely high-ash coal.
'.
6.2 100-hour Tests with Big Brown, Milner, and Black Thunder Coals
t,
Results of the 100-hour tests with Big Brown Coal and Ryton fabric showed that
conditioning significantly reduced particulateemissions and pressure drop. With constant
bag-cleaning frequency and a constant A/C ratio of 4 ft/mln, conditioning resulted in a much
lower baghouse _LP. With an increase in A/C ratio to 6 ft/rain while keeping the bag-
cleaning frequency constant, the _P was lower than the _kPfor the baseline test at an A/C
ratio of 4 ft/min. These results demonstrate the significant benefits of using conditioning
with Big Brown coal and Ryton fabric: a substantialreduction in particulate emissions and a
substantialreduction in pressure drop (or the ability to operate at a higher A/C ratio without
increasing pressure drop or bag-cleaning frequency). The improvements in fabric filter
performance correlate strongly with a shift in the tensile strength-porositycurves and with
increases in the aerated and packed porosity of the fly ash. The residual dust cake with
conditioning was more on the surface of the fabric and somewhat thicker than for the
baseline case. Since there were substantialchanges in fabric filter performance and tensile
strength with conditioning with the 24/12 ppm of NH3/SO3,very likely the conditioning
concentrationscould be reduced somewhat while still retaining a significant benefit.
Conditioning also worked very well with the Milner coal and P84/Ryton fabric,
producing significant reductions in particulateemissions and pressure drop. With a constant
bag-cleaning frequency, the _P with conditioning at an A/C ratio of 6 ft/rain was about the
same as the _P for the baseline case at an A/C ratio of 4 ft/rain. One of the most
encouraging results with the Milner tests was that the residual dust cake weight was very
light for both the baseline and conditioning tests, and there was no evidence of potential bag-
cleaning problems. This indicates that the conditioning concentrations of 24/12 ppm of
NH3/SO 3 were adequate to improve baghouse performance, but not too high to cause bag-
cleaning problems. The conditioning agent concentrationscould likely be increased
somewhat without a negative impact on bag cleanability, possibly resulting in even greater
performance improvement. Similarly, conditioning agent concentrations could be somewhat
reduced, while retaining a significant level of performance improvement.
Conditioning also resulted in some improvements in fabric f'tlterperformance with the

Black Thunder coal, but not as substantial as with the other two coals. Pressure drop was
reduced by about 50% with conditioning at an A/C ratio of 4 ft/rain, but at an A/C ratio of 6
110
ft/min with conditioning, the _P was higher than the ,_P for the baseline test at 4 ft/mim
Particulateemissions were not reduced for the conditioning tests, but emissions were already
t very low for the baseline test with Ryton fabric. Very likely, under other conditions such as
with woven glass fabric, particulate emissions would be somewhat reduced. The shift in the
. fly ash tensile strength curves and the increases in aerated or packed porosity with Black
Thundercoal were smaller than for the other coals. The ammoniaslip was higher for the
• Black Thundertests in spite of injecting only half of the ammonia concentration that was
injected for the other coals. These results,indicate that proper conditioning with Black
Thundercoal may require smaller amounts of conditioning agents than the amounts used for
the other coals. In addition, the ratio of ammoniato SO3may have to be adjustedto
minimize the ammonia slip.
6.3 16..ft/min Tests with Big Brown Coal
For the Big Brown normal dust-loadingtests at an A/C ratio of 16 ft/min and off-line
pulsing, conditioning resulted in a lower K2, but the"e was also some buildupin residual dust
cake weight. This implies that a more energetic pulse may be requiredto sufficiently
dislodge the dust when conditioning is used at very high A/C ratios.
Big Brown tests with a dust loading of approximately 1/20th of normal at 16 fl/min
with off-line pulsing resulted in bag blinding. Neither conditioning nor the use of Neutmlite°
precoating was successful in preventing the bag blinding. The combinationof very low dust
loading and high velocity may allow the dust to pack tightly into the fabric structure,or the
particulate properties may have been different because of different combustion conditions.
6.4 Unresolved Issues and Areas for Future Research
The optimum amountsof conditioning agents to use for aU of the coals are not known.
A 2"1 molar ratio of ammonia to SO3was used for ali of the tests, but it is not known if
optimum performance is always achieved with a 2:1 molar ratio. A research need is to
determine the maximum performance improvement that can be achieved by increasing
conditioning agent concentrations beyond the highest level tested and to better determine a
minimum performance improvement by reducing concentrations below the lowest level
" tested. Since only two concentrations were tested with the Milner coal, a question remains
whether additional performance improvement can be achieved by increasing conditioning
• agent concentrations. One of the difficulties in trying to establish optimum performance
improvement is that the characteristics of an ideal dust that provide the best combination of
111
good pore-bridging ability, high porosity, and good bag cleanability are not known. A much
better understandingof the relationships between the cohesive properties of dusts and
filtration performance is needed before conditioning can be refined for optimum performance
improvement.
The results clearly show the substantialbenefits in fabric f'flterperformance that can be
achieved with conditioning. Longer-term testing is needed, however, to establish the effect t
of conditioning on the residual dust cake over an extended period. The chemistry of the
process also needs further investigation so that possible chemical reactions between the
conditioning agents and fly ash or the conditioning agents and the fabric could be predicted.
Any potential effect on long-term fabric degradationalso needs to be assessed.
The results from the 16-ft/rain tests with a reduced dust loading were difficult to
explain. More work is needed to determineif the combinationof a high face velocity and a
low dust loading is more likely to cause fabric blinding than a normal dust loading and high
. velocity, or if cofiring with coal and naturalgas alters particulateproperties to facilitate
fabric blinding. In addition, it is not known whether conditioning would be effective with a
very low dust loading by using different co_nditioningagent concentrations.
A question exists as to the effect, ff any, that small amounts of ammonium sulfate may
have on fly ash disposal and u"tflizafion.The concentmfiou of ammonia on the fly ash was
measuredto be 0.1% to 0.3 % (most likely in the form of ammonium sulfate). Three
potential concerns are possible off-gassing of ammonia upon wetting of the fly ash, leaching
characteristics of the ammonia m' ammoniaproducts, and the effect on fly ash utilization.
In an aqueous solution at a pH greeter than 10, the ammonia is in the form of dissolved
ammonia gas and would likely result in some off-gassing of ammonia into the air.
Therefore, in the case of wetting of the fly ash or under very high-humidity conditions with
high-calcium or high-sodium ashes, the possibility of a detectable ammonia odor exists.
With dry fly ash handling and disposal, ammoniaodor is not expected to be a problem.
ASTM leachate tests were not part of this study but have been conducted on fly ashes
from previous baseline and conditioning tests. As expected, leachate concentrations of
NH4 + , SO4 2", and NO3"were higher for the conditioned fly ashes. The NO3" is apparently the
result of oxidation of some of the ammonia. Nitrate in groundwater is regulated by the
primary drinking water standards;however, the contribution of nitrate from conditioned fly
ash to overall nitrates is likely to be very small compared to major contributors such as
112
agriculturalsources. Soluble sulfates are typically found in fly ash so the addition of a small
amountof soluble SO_ should not change disposal considerations. NI_ + is not normally
found in fly ash, but apparentlythere are no regulations that would require special disposal
methods for fly ash containing small amounts of ammonia. One interesting result of the
• leaching tests was that selenium concentrationin the leachatewas reduced by about 75%
which implies that conditioning inhibits the leaching of selenium. The mechanism, however,
' is not clear, and this result needs to be verified.
Fly ashes have been used extensively for cement replacement in concrete. The effect
of the ammonia on the pozzolonic activity of the fly ash is not known. The concentration
and chemical form of the ammonia on the fly ash and the specific use will likely determine
whether small amounts of ammoniaon fly ash will significantly affect utilization. Research
is needed on the effect of ammoniaon fly ash disposal and utilization, not only for flue gas
conditioning applied to fabric filtration, but also for other processes such as NOx control by
selective catalytic reductionwhich result in ammonia on fly ash.
7,0 REFERENCES
1. Miller, S.J.; Laudal, D.L. "Rml-TimeMeasurementof Respirable Particulate

Emissions from a Fabric Filter," In Particulate and Multiphase Processes, Vol. 2,
Contamination Analysis and Control; Ariman, T., Ed.; Hemisphere Pub. Corp., 1987,
p 663.
2. Miller, S.J.; Laudal, D.L. "ParticulateRemoval Enhancementof a Fabric Filter Using

Flue Gas Conditioning," Presented at the ThirdEPRI Conference on Fabric Filter
Technology for Coal-FiredPower Plants, Scottsdale, AZ, Nov. 19-21, 1985.
3. Laudal, D.L.; Miller, S.J. "Flue Gas Conditioningfor Improved Baghouse

Performance," In Proceedings of the Sixth Symposium on the Transfer and Utilization
of Particulate Control Technology; EPRI CS-4918, Nov. 1986, Vol. 3, p 14-1.
4. Miller, S.J.; Laudal, D.L. "Flue Gas Conditioning for Improved Fine Particle Capture
in Fabric Filters: ComparativeTechnical and Economic Assessment," In Low-Rank
Coal Research Final Report, Vol. II, Advanced Research and Technology Development;
DOE/FC/10637-2414 (DE87006532), Apr. 1987.
5. Miller, S.J.; Laudal, D.L. "Fine Particulate Emissions: Flue Gas Conditioning for
Improved Fine-Particle Capture in Fabric Filters," Final Technical Reportfor the
Period Apr. 1, 1987, through Mar. 31, 1988; DE-FC21-86MCI0637, GrandForks,
• ND, August 1988.
113
6. Laudal, D.L.; Miller, S.J. "Flue Gas Conditioning for Baghouse Performance
Improvementwith Low-Rank Coals," In Proceedings of the Fourteenth Biennial Lignite
Symposium on the Technology and Utilization of Low-Rank Coals; University of North
Dakota Energy and EnvironmentalResearch Center, GrandForks, ND, 1987.
7. Miller, S.J.; Laudal, D.L.; Kim, S.S. "Mechanisms of Fabric Filter Performance
Improvement with Flue Gas Conditioning," In Proceedings of the Seventh EPA/EPRI
Symposium on the Transfer and Utilization of ParticulateControl Technology; EPRI
GS-6208, Feb. 1989, Vol. 2, pp 25-1.
8. Miller, S.J. "Flue Gas Conditioning for Fabric Filter Performance Improvement,"
Final Project Report for ContractNo. DE-AC22-88PC88866 for Pittsburgh Energy
Technology Center, Dec. 1989.
9. Miller, S.J.; Laudal, D.L. "EnhancingBaghouse Performance with Conditioning

Agents: Basis, Developments, and Economics, wIn Proceedings of the Eighth
Particulate Control Symposium; EPRI GS-7050, Nov. 1990, Vol. 2, p 23-1.
10. Miller, S.J.; Laudal, D.L; Chang, R.L. WFlueGas Conditioning for Improving Pulse-
Jet Baghouse Performance, _ Presented at the Ninth ParticulateControl Symposium,
Williamsburg, VA, Oct. 15-18, 1991.
11. Miller, S.J.; Heidt, M.K.; Laudal, D.L.; Weber, G.F. "Flue Gas Cleanup,_
Semiannual Report for the Period Jan. 1, 1992 - Jun. 30, 1992 for ContractNo. DE-'
FC21-86MC10637, Jul. 1992.
114

Baghouse

Uploaded by

Copyright:

Available Formats

Baghouse

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Baghouse

Uploaded by

Copyright:

Available Formats

DOE/MC/lO637-3312

Work Performed Under Contract No.: DE.FC21-86MC10637

A pilot study was conducted at the Energy and EnvironmentalResearch Center

LIST OF TABLES ..................... ix

• 1.0 PROJECT SUMMARY ..................................... l

2.0 PROJECT OBJECTIVES .................................... 4

4.0 FACIIXrY DF,,SCRIFFION AND PROCEDURF.S .................... 99

6.0 CONCLUSIONS AND DISCUSSION ............................ 99

3 Schematic of the paxticulate-samplingsystem (APS 33, DM:PS, and dilution

4 Schematic of the Cohetester ................... ............. 16

5 Log-probability plot of coal particle-size distributionbased on sieve

9 Respirable mass emissions and pressuredrop as a function of time for Big

I0 Pressuredropforbaseline andconditioning tests withBigBrown coalwith

16 Comparison of submicron emissions measured with the CNC for Big

18 Respirable mass emissions and pressure drop as a function of time for

19 Respirable mass emissions and pressure drop as a function of time for

21 Respirable mass emissions and pressure drop as a function of dme for

22 Respirable mass emissions and pressure drop as a function of time for

23 Respirable mass emissions and pressure drop as a functionof time for

24 Particle-sizedistributionof Big BrownandMilner fly ashbasedon

25 Tensilestrengthasa functionof porosityfor baselineandconditionedfly

27 Tensile strength as a function of porosity for Task 2 Milner tests ......... 48

28 Effect of conditioning on baghouse pressure drop for 100-hour Big Brown

29 Effect of conditioning on average respirablemass emissions for 100-hour

34 Range of submicron particle emissions for the 100-hour conditioning

35 Range of submicron particle emissions for the 100-hour conditioning

36 Effect of conditioning on baghouse pressure drop for 100-hour Milner tests

37 Effect of conditioning on average respirable mass emissions for 100-hour

41 Range of submicronparticle emissions for the 100--hourbaseline Run 406

42 Range of submicron particle emissions for the 100-hour conditioning

43 Range of submicron particle emissions for the 100-no,,Tconditioning

44 Effect of conditioning on baghouse pressure drop for 100-hour Black Thunder

45 Effect of conditioning on baghouse pressure drop for Black Thunder test

46 Effect of conditioning on average respirable mass emissions for 100-hour

91 Range of submicron particle emissions t'or the 100-hour conditioning

52 Range of submicron particle emissions for the 100-hour conditioning

54 Range of submicron particle emissions for Black Thunder test with

55 Effect of conditioning on baghouse pressure drop for 8-hour Big Brown

56 Effect of conditioning on average respirable mass emissions for 8-hour Big

61 Particle-size distribution of Big Brown fly ash based on multicyclone sampling . 82

62 Particle-size distribution of Milner fly ash based on multicyclone sampling .... 82

63 Comparison of particle-size distributions of Big Brown, Milner, and Black

64 Comparison of particle-size distributions for normal and reduced dust-

67 Effect of conditioning on fly ash tensile strength for 100-hour Black

68 Comparison of fly ash tensile strengthsfor conditioned tests with Big

73 Laboratory measurement of K: as a function of porosity with Carman-

74 Laboratory measurement of K, as a function of porosity with Carman-

75 Laboratory measurement of K: as a function of porosity with Carman-

77 Correlation between initial K2 and aerated or packed porosity 95

78 Correlationbetween effective K2and aerated or packed porosity .......... 95

83 Residual dustcake at the end of the 100-hourconditioning test with Milner

I Matrix for Task 2 ...................................... 7

2 Matrix for Task 3 ...................................... 8

3 Pulse-Jet Fabric Filter Specifications ........................... I1

4 ACGIH Respirable Mass Del'tuition ............................ 13

5 Summary of Coal Analyses ................................. 19

5 Results of Task 2 Testing .................................. 23

7 Physical and Chemical Fly Ash Characteristics from Task 2 ............. 24

8 Results from Task 3 Testing ................................ 51

9 Physical and Chemical Fly Ash Characteristics from Task 3 ............. 52

. .._ ,o.l* _' : ....... _L_