United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-81-105 Dec. 1981
Project Summary
Control of Particulate
Emissions from Atmospheric
Fluidized-Bed Combustion
With Fabric Filters and
Electrostatic Precipitators
David V. Bubenick, Robert R. Hall, and John A. Dirgo
Fabric filters are being installed on
many new atmospheric fluidized-bed
combustion (AFBC) units, despite the
lack of test or demonstration data. For
this reason, the present study focuses
on assessing fundamental chemical
and physical characteristics affecting
the performance of particulate control
equipment based on five fly ash
samples from full- and pilot-scale
AFBC units. These results were used
in conjunction with fabric filter (FF)
and electrostatic precipitator (ESP)
mathematical models to illustrate how
control device performance may be
affected by AFBC fly ash properties.
Laboratory measurements of the
specific resistance coefficient (K2), a
measure of the pressure loss through a
dust deposit on a fabric, ranged from
1.9 to 5.3 N-min/g-m for the AFBC
fly ashes, compared to values near 2.0
N-min/g-m for conventional fly ash.
Mathematical simulation of an opera-
ting FF indicated that an increase in K2
from 2 to 5 N-min/g-m could result in
a 30 percent increase in penetration
and a doubling of pressure loss. Low
electrical resistivity values (105 to 107
ohm-cm) of two fly ash samples
containing high carbon (34 to 46
percent) would be expected to cause
poor ESP performance. The remaining
three samples exhibited resistivities
falling within the range of conventional
combustion design experience with
ESPs as shown by model simulation.
Efficient operation of the mechanical
precollectors or combustion efficiency
improvements may reduce the impact
of high carbon carryover and thereby
minimize possible fabric blinding for
FFs and sparkover resulting in poor
ESP performance. The ultimate choice
of a particulate control device must
take into account all of the above
factors as well as case-by-case and
transient AFBC operating conditions.
This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory, Research Tri-
angle Park. NC. to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
As part of an engineering assessment
of control technology for atmospheric
fluidized-bed combustion (AFBC), EPA's
Industrial Environmental Research
Laboratory at Research Triangle Park,
NC, is investigating the suitability of
fabric filters and electrostatic precipi-
tators (ESPs) as final particulate collec-
tors. Initial efforts were directed at
examining selected planned and oper-
-------
ating fabric filters and ESPs applied to
AFBC boilers. This survey led to the
conclusion that fabric filters are being
installed on many new AFBC units
despite the lack of comprehensive test
or demonstration data. ESPs, on the
other hand, have been applied on a
rather limited basis. With this lack of a
working data base, a more fundamental
approach was elected to assess panicu-
late control device suitability. This
consisted of first making basic mea-
surements on selected fly ash samples
obtained from operating pilot- and full-
scale AFBCs in the U.S. Basic relation-
ships between key design and perform-
ance parameters for fabric filters and
ESPs were examined using analytical
mathematical performance models. The
results of measurements and relation-
ships are presented in this report, along
with an assessment of additional data
needs.
Approach
Five fly ash samples were obtained
from fabric filter hoppers at four AFBC
sites ranging in size from pilot- to full-
scale units. A series of fundamental
laboratory measurements were per-
formed on each sample to characterize
particle size distribution, specific resis-
tance coefficient (a measure of fly ash
filterability), electrical resistivity, chem-
ical composition, carbon content, and
morphology, as shown in Table 1. Test
results were then used in conjunction
with fabric filter and ESP mathematical
models to illustrate how control device
performance may be affected by AFBC
fly ash properties.
The fabric filter mathematical com-
puter model, developed by GCA/Tech-
nology Division for EPA, can be used to
predict the field performance of fabric
filter systems applied to coal-fired utility
boilers.112'3 The model applies to specific
collector types using woven glass
fabrics and simulates fabric cleaning by
a collapse and reverse flow process
and/or shake assist. Required operating
parameters include flue gas tempera-
ture, inlet dust concentration, and gross
filtration velocity (air-to-cloth ratio).
Among the important dust properties to
be considered is the specific resistance
coefficient of the dust, K2. This param-
eter describes the linear increase in
resistance to air flow that occurs when
dust deposition on the fabric filter bags
is uniform over time. In addition, an
estimate of the effectiveness of the
fabric cleaning process is required. This
estimate is provided by ac, the ratio of
the mass of dust removed from a bag
during the cleaning cycle to the mass of
dust on the bag prior to cleaning. In this
study, the effects of K2 on pressure drop
and penetration, for different levels of
filter face velocity and ac, were examined.
The ESP computer model, developed
by Southern Research Institute (SoRI)
for EPA, simulates ESP operation and
performance characteristics by incor-
porating fundamental theoretical rela-
tionships that describe the physical,
chemical, and electrostatic mechanisms
interacting in the ESP process.4'5
Nonideal effects such as nonuniform
gas velocity distribution, gas bypassage
of electrified regions, and particle reen-
trainment due to rapping and other
causes are accounted for by empirical
correction factors. The SoRI/EPA model
was used in this study to investigate the
relationship between specific collection
area (SCA) and overall mass collection
efficiency using laboratory measured
resistivity and particle size distribution
data.
Results/Discussion
Table 2 shows the results of resistivity,
particle chemical composition, particle
size distribution, and filtration mea-
surements of the five fly ash samples.
The specific resistance coefficient, K2,
of the FBC fly ash samples varies from
1.94 to 5.25 N-min/g-m (11.6 to 31.4
in. W.C.-min-ft/lb). Filtration theory
predicts that Kz increases with decreasing
particle size. A specific surface param-
eter, S02, calculated from the mass
median diameter and geometric standard
deviation of a dust sample, may be used
to characterize this relationship between
K2 and particle size.1 The computed
values of S02 in Table 2 appear to be
correlated with measured K2 (correla-
tion coefficient r = 0.73). The actual
point scatter may be attributed to
differences in particle shape, surface
characteristics, or charge effects,
although additional investigation is
needed to resolve this issue.
The GCA fabric filter model was used
to examine the effects of variations in K2
on filter penetration and pressure drop.
An average fabric filter system was
defined, and K2 was chosen to be 5.0 N-
min/g-m (30 in. W.C.-min-ft./lb) inf
contrast to 2.0 N-min/g-m (12 in. W.C.-"
min-ft/lb), a value which is reported for
conventional coal-fired boiler fly ash.1
Figures 1 and 2 show the effect of K2
on penetration and pressure drop,
respectively, as a function of filter face
velocity. At an air-to-cloth ratio of 0.61
m/min (2 ft/min), the difference in K2
results in roughly a 30 percent increase
in fractional penetration and a 100
percent increase in pressure drop. To
Table 1. Summary of FBC Fly Ash Measurements
Parameter/analysis
Instrument/technique
Particle size distribution
Specific resistance
coefficient
Resistivity
Chemical/spectral analysis
Carbon analysis
Morphology
Andersen Mark III (eight-stage) cascade impactor
GCA/Technology Division bench-scale fabric filter test panel system
Denver Research Institute single- and multiple-cell resistivity apparatus
Amray Model 1200 SEM* with EDXRA*
Low-temperature ashing (LOIf
High-temperature ashing (CHN)"
Amray Model 12OO SEM*
"Scanning Electron Microscope.
"Energy Dispersive X-ray Analysis unit.
cLoss on Ignition.
''Carbon-Hydrogen-Nitrogen test using a Per kin Elmer Model 240B Elemental Analyzer.
-------
Table 2. Results of Fundamental Measurements Performed on FBC Fly Ash Samples
Resistivity fohm-cm)
@ 70% moisture
Chemical/ spectral analysis'
constituents (% by weight)
Sample designation
149°C 260°C 370°C AI203 S/02 K,0 CaO Ti02.
Particle size Specific surface
distribution parameter Specific resistance
So2 (x 70"°^" coefficient, K2
aMMD. x
(urn) <7g (cm *) (N-min/g-m)
A" ' Alliance
B - SATR
C" - FluiDyne
D - FluiDyne
E - Georgetown
3.3 xlO7 1.2x10* ' 11.1
3.2 x 10" 5.0 x 7010 5.0 x 70s 27.4
8.0 x 10" 1.4 x 10" 5.0 x 10* 12.0
1 1 x 70'2 2.6 x 70" 8.0 x 10' 72.3
7.;*;o! 3.3 x 10s ' 99
33.2
48.5
39.5
40.0
24.2
1.8
1.7
3.0
3.0
1.3
10.1
107
20.7
22.6
7.0
1.0
1.6
1.8
1.0
—
4.1
3.6
7.8
7.6
6.4
33.7
7.5
102
9.5
46.2
6.0
53
7.3
8.9
7.7
2.8
26
26
30
2.7
5.77
6.39
3.37
3.04
3.26
4.89
5.25
7.34
2.46
4.74
"The six oxides were derived from SEM/EDXR analysis, a method relatively insensitive to elements of low atomic number. Some elements (attributed to their common
oxides; e.g., SOii could not be measured; hence the total does not equal 100%.
"The specific surface parameter, S02, can be calculated from the particle size distribution parameters by the formula
So2 =
T / ,07.757/0gVg \ "I
LI MMD )J
Particle density, used to convert aMMD to MMD, was assumed to be 2.0 g/cm3.
'Percent carbon was determined using a CHN test. Results closely agree with those from a simple LOf determination For samples A, B, C, D, and E the percent
combustibles fLOIs) were 25.9, 7.5. 12.4. 10.9. and 43 0, respectively.
"Without fly ash reinfection.
'Both samples A and E sparked over above 260°C (500°F), hence, testing could not be continued.
achieve a collection efficiency of 99.5
percent, for example, a 0.91 m/min (3
ft/min) filtering velocity would be
acceptable with a fly ash having the
lower K2, but not with a fly ash at the
higher K2 level.
If it is found that FBC fly ash does not
have dust-fabric release properties
similar to conventional fly ash, then the
leanability of the fabric would be
expected to be different. For a K2 of 5.0
N-min/g-m (30 in. W.C.-min-ft/lb) and
an air-to-cloth ratio of 0.61 m/min (2
ft/min), the model calculates that
approximately 25 percent of the fly ash
is removed by cleaning. As the degree of
cleaning increases from 10 to 40
percent, pressure drop is reduced by
approximately 70 percent while pene-
tration increases by roughly 12 percent.
Clearly, the ease of dust removal from
the bag surface as a result of cleaning is
an important factor in predicting fabric
filter operating pressure drop and
performance.
Particle resistivity is an important
consideration in the application of ESPs
to FBC. Very low concentrations of S03
have been recorded in FBC flue gas,
which implies a reduced conditioning
effect on the particles, resulting in
increased resistivity. In addition, all
sorbent materials (CaCO3, CaO, MgO,
and CaSOJ have high resistivities.
However, carbon content exerts a
strong influence in lowering fly ash
resistivity.
Samples B, C, and D exhibited
resistivities in the moderate to high
range generally observed with conven-
?nal coal-fired boiler fly ash. The
characteristic variation of resistivity
with temperature was also observed.
The unusually low resistivities for
samples A and E may be attributed to
the high carbon contents present. These
resistivities were similar to results
obtained from a test run conducted on a
hot ESP at the Rivesville, WV, FBC,
where the carbon content was 47
percent.8 In view of the conditions that
promote high carbon carryover in an
AFBC unit, it may be a more significant
operating problem than in conventional
combustion. This emphasizes the
importance of the primary cyclone in
providing reliable, high efficiency
paniculate collection capability so that
unburned carbon is returned to the
boiler for efficient combustion and the
downstream collector is protected from
the effects of carbon carryover.
Figure 3 shows the effect of fly ash
resistivity on cold ESP overall mass
collection efficiency as a function of
specific collection area. The slope of the
reference line reflects the exponential
design relationship between SCA and
collection efficiency. The SoRI ESP
model was used to predict the col lection
efficiencies at an SCA of approximately
49.3 mz/(mVsec), (250ftz/1000acfm),
for samples B, C, and D using their
measured resistivities at 149°C (300°F)
and particle size distributions. These
results were extrapolated over a range
of SCAs using the slope of the reference
line shown by the cross-hatched band. It
is noted that the predicted SCA re-
quirements for samples B, C, and D,
with resistivities between 1011 and 1012
ohm-cm, fall within the range of design
experience for cold ESPs applied to the
collection of conventional fly ash.
The SoRI model was also used to
investigate particle size effects on ESP
performance. The use of cyclones for fly
ash recycle in FBC systems may result
in finer particle sizes, relative to
conventional fly ash, reaching the ESP.
Model results show that the collection
of smaller particles requires a larger
ESP for a given efficiency, if all other
factors are held constant. Because the
cyclones also reduce the ESP inlet
loading, the smaller particle size
distribution of FBC fly ash does not
necessarily imply a higher SCA re-
quirement. However, depending on the
collection efficiencies of the cyclones,
as well as other important factors such
as rapping reentrainment losses, a
more conservative estimate of SCA may
be required.
Field Operating Considerations
Some of the more common problem
areas encountered with fabric filter
operation include fabric/dust deposit
interaction, bag failures, design or
maintenance failures relating to bag
support and the cleaning mechanism,
and structural design problems including
isolation dampers and the ash removal
system. For example, a dust which is
hygroscopic may, in the presence of
moisture, cause blinding, impede proper
cleaning, cause excessive pressure
buildup across the fabric, reduce bag
service life, and promote ash removal
problems.
Potential problem area considerations
for ESPs applied to FBC units include:
-------
0.03
0.02
0.30
Air-to-Cloth Ratio (ft/min)
0.61 0.91 1.22
1.52
0.01
c
S 0.007
ra
i_
05
c
- 0.005
ID
C
g
o
CD
0.003
0.002
0.007
I I I
Average System Definition
No. of Compartments 14
Cleaning Cycle Time (min) 35
Gas Temperature (°C) 205
I
Inlet Dust Cone. (g/dsmj)
8.0
FBCFIyAsh
— 5.0 N-min/g-m
Conventional
Fly Ash
K2 = 2.0 N-min/g-m
_L
37.0
98.0
99.0
33.3 =f.
33.5 3>'
39.7
99.8
99.9
234
Air-to-Cloth Ratio (m/min)
and operation to handle most of the
potential transient conditions in AFBC
operation.
References
1.
4.
6.
Dennis, R., et al. Filtration Model for
Coal Fly Ash with Glass Fabrics.
GCA/Technology Division. EPA-
600/7-77-084 (NTIS PB 276489).
August 1977.
Dennis, R., and H.A. Klemm. Fabric
Filter Model Format Change: Volume
I. Detailed Technical Report. GCA/
Technology Division. EPA-600/7-
79-043a (NTIS PB 293551). February
1979.
Dennis, R., and H.A. Klemm. Fabric
Filter Model Format Change: Volume
II. User's Guide. GCA/Technology
Division. EPA-600/7-79-043b(NTIS
PB 294042). February 1979.
McDonald, J.R. A Mathematical
Model of Electrostatic Precipitation
(Revision 1): Volume I. Modeling and
Programming. Southern Research
Institute. EPA 600/7-78-111 a (NTIS
PB 284614). June 1978.
McDonald, J.R. A Mathematical
Model of Electrostatic Precipitation
(Revision 1): Volume II. User Manual.
Southern Research Institute. EPA-
600/7-78-111 b (NTIS PB 284615).
June 1978. f
Pope, Evans, and Robbins, Inc.™
Multicell Fluidized-Bed Boiler Design,
Construction and Test Program.
Research and Development Report
No. 90, Interim Report No. 1 for
Period October 1972-June 1974.
Prepared for the U.S. Department of
the Interior. PB 236 254. August
1974.
Figure 1. The effect of air-to-cloth ratio on particle penetration for different
levels.
electrode fouling due to buildup of fly
ash on the wires; local burning on the
wires due to high carbon carryover (and
manifested by heavy sparking); rapping
reentrainment losses; and transient
conditions associated with boiler startup.
The performance of the final panicu-
late collector depends on the FBC
operating parameters; in particular on
cyclone efficiency, carbon burnup cell
combustion efficiency, optimized free-
board height, bed depth, and fluidizing
velocity. The carryover of unburned
carbon during startup or during a
transient boiler operating condition may
severely deteriorate electrical conditions
in an ESP, with the possibility of ignition
and fire. In fabric filters, this same
condition may cause fabric blinding and
permanent damage. Low load conditions
can reduce cyclone efficiency and add to
the carbon carryover problem. One
important conclusion of this report is
that the precollector or cyclone must be
considered as an integral part of the
total particulate control system. Fur-
thermore, the entire collector system
must be flexible enough in both design
-------
to
-------
°o rs rx
io is' oj
Specific Collection Area (m2/(m3/s))
05
99.9
99.8
99.0
98.0
1 95.0
.QJ
.5
03
fx
qi ».
Uj
90.0
85.0
80.0
70.0
§
Conventional Combustion
D Lignite Coal
Bituminous Coal
Fluidized Bed Combustion
//// Samples B, C.D
\ [ [\\i\iii
Figure 3.
§O O OOOOOOOOO
o o ooooooooo
Specific Collection Area (ft2/WOO acfm)
Effect of fly ash resistivity on cold ESP overall mass collection efficiency
as a function of specific collection area.
D. V. Bubenick, R. R. Hall, and J. A. Dirgo are with GCA/Technology Division.
Burlington Road, Bedford, MA 01730.
John O. Mill/ken is the EPA Project Officer (see below).
The complete report, entitled "Control of Paniculate Emissions from Atmos-
pheric Fluidized-Bed Combustion with Fabric Filters and Electrostatic Precip-
itators." {Order No. PB 82-115 528; Cost: $10.50, subject to change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
U.S. GOVERNMENT PRINTING OFFICE:1981--559-092/3361
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
PS 0000329
U S tNVIR PROTECTION AGENCY
REi*IONi 5 LIBRARY
ajO S DEARbGKN STKEEJ
CHICAGO IL 60604
------- |