EPA-650/2-75-033
March 1975
Environmental Protection Technology Series
I
55
\
ssE
UJ
a
-------
EPA-650/2-75-033
PARTICIPATE COLLECTION
EFFICIENCY MEASUREMENTS
ON A WET ELECTROSTATIC
PRECIPITATOR
by
John P. Gooch and Joseph D. McCain
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
for
TheM. W. Kellogg Co.
1300 Three Greenway Plaza
Houston, Texas 77046
Contract No. 68-02-1308, Task 21
ROAP No. 21ADL-004
Program Element No. 1AB012
EPA Project Officer: Leslie E. Sparks
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON, D. C. 20460
March 1975
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have* been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOC1OECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to
develop and demonstrate instrumentation , equipment and methodology
to repair or prevent environmental degradation from point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA-650/2-75-033
ii
-------
ABSTRACT
Fractional and overall particulate collection efficiency
measurements were made on a plate-type wet electrostatic
precipitator collecting fume from an aluminum pot line.
»•
Overall collection efficiency determinations, based on a
mass train with an in-stack filter, ranged from 95.0 to
98.0%. The mass filter obtained much higher total outlet
mass loadings than did the Andersen impactors, presumably
because of large entrained liquor droplets which were
captured by the mass traverse, but not by the single-point
impactor measurements. The average minimum collection
efficiency in the size range 0.2 to 1.0 ym diameter
(based on the Andersen data) was 98.5%.
Comparisons between measured (with Andersen impactors) and
predicted collection efficiencies obtained from a mathe-
matical model of an electrostatic precipitator indicated
fair agreement in the size range from 0.2 to about 1.3 ym.
For larger particles, the collection efficiency-particle
size relationship departed drastically from the expected
pattern, possibly because of liquor carryover from the
electrode irrigation system.
111
-------
CONTENTS
?age
Abstract iii
List of Figures v
List of Tables vi
Conversion Factors vii
Sections
I Summary and Conclusions 1
II Introduction 3
III Measurement Techniques 7
IV Results 14
V Discussion 36
VI References 48
VII Appendix 50
IV
-------
FIGURES
No. Page
1 Schematic of scrubber-precipitator system and
sampling locations 4
2 Schematic of primary emission control system at
the test site 5
3 Comparison of sedimentation and equivalent
optical diameters 9
4 Optical and diffusional sizing system 10
5 Inlet Andersen data 17
6 Outlet Andersen data 18
7 Andersen data on log probability co-ordinates 19
8 Comparison of Andersen and Brink average inlet
data 20
9 Inlet differential size distribution 22
10 Outlet differential size distribution 23
11 Andersen fractional efficiency data 24
12 Relative concentration variation from condensa-
tion nuclei counter 26
13 Relative concentration variation from optical
particle counter 27
14 Cumulative size distributions on a number basis for
various industrial particulate sources as measured
by optical and diffusional methods 28
15 Measured fractional efficiencies for a wet elec-
trostatic precipitator with the operating parameters
as indicated, installed downstream of a spray-type
scrubber on an aluminum reduction pot line 29
16 Wet electrostatic precipitator 37
17 Schematic of electrode arrangement 38
18 Voltage current relationship and operating ranges 41
v
-------
TABLES
No. Page
1 Weight changes of Andersen substrate after
sampling filtered effluent from wet ESP 12
2 Andersen inlet data 15
3 Andersen outlet data 16
4 Mass train test results 31
5 Results from tests conducted by a local
pollution control agency 33
6 Summary of specifications for the wet electro-
static precipitators 40
7 Effect of dielectric constant on predicted
efficiencies 43
8 Precipitation rate parameters 45
9 Qualitative comparison of mass measurements 45
10 Operating power estimated for wet electro-
static precipitator no. 553 47
VI
-------
CONVERSION FACTORS
To Convert From
Ibs
grains/cf
cfm
lbs/in2
OF
ft2/1000 cfm
inches w.g.
gallon
ft
inches
To
Kg
grams/m3
m3/sec
Kg/m2
°C
m2/(m3/sec)
nun Hg
liter
m
m
Multiply By
0.454
2.29
0.000472
703
(°F - 32) x 5/9
0.197
1.868
3.785
0.3048
0.0254
via.
-------
SECTION I
SUMMARY AND CONCLUSIONS
Fractional and overall particulate collection efficiency
measurements were made on a plate-type wet electrostatic
precipitator collecting fume from an aluminum pot line con-
sisting of horizontal stud self-baking aluminum reduction
cells. The average mass median diameter of the inlet particu-
late, as determined with Andersen impactors, was 0.60 ym, and
the average minimum collection efficiency in the size range
0.2-1.0 ym diameter (again based on Andersen data) was 98.5%.
Overall collection efficiency determinations, based on a mass
train with an in-stack filter, ranged from 95.0 to 98.0%.
The mass filter obtained much higher total outlet mass loadings
than did the Andersen impactors (4.2 mg/am3 vs 0.84 mg/am3),
presumably because of large entrained liquor droplets which
were captured by the mass traverse, but not by the single-
point impactor measurements. Agreement was obtained between
total mass loadings from the Andersens and the mass filter at
the inlet.
Diffusional measurements indicated an increasing collection
efficiency in the size range from 0.1 to 0.07 ym, with
an indicated collection efficiency of 99.6% for 0.07 ym
diameter particles. Optically and inertially determined
collection efficiencies showed fair agreement in the size
range 0.3-1.0 ym, but the conversion of the optical number
density to mass loadings showed that the optical instrument
was detecting at least one order of magnitude less mass than
were the Andersen impactors at equivalent particle sizes, at
both the precipitator inlet and outlet. Possible reasons for
the discrepancy are evaporation of volatile components in
the optical system, loss of particle count due to systematic
errors in the sampling system and particle counter, and the
steep slope of the size distribution on a number basis.
-------
The estimated operating power requirements for the wet
precipitator totaled 114 kw, or 2.58 kw/(m3/sec). This
includes high voltage power requirements, pumping power, fan
power, based on 1.27 cm (1/2") water column pressure drop,
and power to heat the high voltage insulators. If electri-
cal power costs are $0.01/kwh, the energy costs for operating
the unit are estimated as $27.00/day.
Comparisons between measured (with Andersen impactors) and
predicted collection efficiencies obtained from a mathematical
model of an electrostatic precipitator indicated fair agree-
ment in the size range from 0.2 to about 1.3 ym. For larger
particles, the collection efficiency-particle size relation-
ship departed drastically from the expected functional form,
presumably because of liquor-carryover from the electrode
irrigation system.
-------
SECTION II
INTRODUCTION
This report presents the results obtained from a performance
test conducted by Southern Research Institute on a wet elec-
trostatic precipitator collecting fume from horizontal stud
Soderburg aluminum reduction cells during the period
August 19-23, 1974. The wet electrostatic precipitator
was designated as ESP No. 553. The objectives of the test
series were (1) to determine the fractional and overall
particulate collection efficiency of the electrostatic
precipitator, (2) to compare the measured performance of
the precipitator with that projected from a mathematical
model.
At the reduction plant, wet precipitators are installed
both with and without a spray tower prior to the precipita-
tor. Since the occurrence of condensation within the precipi-
tator itself confuses interpretation of the data, it was
decided to conduct the test series on a unit which is preceded
by two spray towers, to minimize this effect. The spray
towers treat exhaust gas from 28 pots with an alkaline solu-
tion which cools the gas from about 220°F to about 100°F.
Figure 1 shows the arrangement of the wet precipitator,
scrubbers, and sampling locations. Figure 2 is a schematic
of the liquor flow through the system given by Bakke.1
Horizontal stud Soderburg cells are of the "self-baking"
type, in that the carbon electrode is baked within the
cell. The effluent from the cell therefore contains hydro-
carbons volatilized from the binders used to make the anode.
Other constituents include HF gas, which results from hydro-
lysis of fluoride salts, particulates of vaporized bath
-------
TOP OF HUT MAY BE USED
TWO HOOKS FOR STOPS
SCRUBBERS
PLATFORM,
VERT
SAMPLING
HUT, HORIZ. SAMPLING
WET
•ELECTROSTATIC
PRECIPITATOR
Figure 1. Schematic of scrubber-precipitator system and sampling locations
-------
STACK
3 T-R SETS
POT GAS MANIFOLD
LIQUOR.
MAIN
-&i
a a
WEP SPRAYS
INLET DUCT
RECEIVING
.TANK
SCRUBBER
SPRAYS
MAIN FAN
BOOSTER
PUMP
CYCLONIC
SCRUBBERS
(TWO)
LIQUOR
RETURN
Figure 2. Schematic of primary emission control system1
-------
materials, and alumina, cryolite, and other dusts entrained
from the bath crust. During the course of the test period,
routine operations were in progress on the cells which
supply the scrubber-precipitator system. These operations
include the breaking of the crust in each cell at approximately
2 hr intervals, and anode maintenance operations known as
"pin and channel pulls" and "flex raises". The anode
maintenance and the crust breaking is performed on the cells
on an individual basis. Thus the effect of the individual
operations on the total particulate concentration entering
the wet precipitator is somewhat damped. Table A-l in the
Appendix gives the operations performed during the test
period on the cells which are vented to the wet ESP.
-------
SECTION III
MEASUREMENT TECHNIQUES
PARTICLE SIZE MEASUREMENTS
Particle size and concentration measurements were conducted
using the following methods: (1) diffusional techniques using
condensation nuclei counters and diffusion batteries for
determining concentration and size distribution on a number
basis for particles having diameters less than approximately
0.2 ym, (2) optical techniques for determining concentra-
tions and size distributions for particles having diameters
between approximately 0.3 ym and 1.5 ym, (3) inertial techniques
using cascade impactors for determining concentrations and
size distributions on a mass basis for particles having
diameters between approximately 0.25 ym and 5.0 ym. A
detailed description of these measurement techniques is
given elsewhere,2 and therefore only a brief discussion will
be given in this report.
For optical and diffusional measurements, extensive dilution
of the gas stream being sampled is usually required because
of the limitations imposed by the useful ranges of both the
optical counter and condensation nuclei counter. Dilution
ratios ranging from 0 to 20 were used at the outlet, and
from 30 to 90 at the inlet. As a general practice, checks
of the linearity of particle count with dilution changes
are performed to determine whether any anomalies resulting
from condensation or other phenomena are occurring within the
measurement system.
Due to limitations imposed by equipment availability, it was
not possible to obtain simultaneous measurements at the
precipitator inlet and outlet with the optical and diffusional
instruments. However, the particulate concentrations were
sufficiently stable to enable meaningful fractional efficiency
7
-------
data to be derived by first obtaining inlet data, and
subsequently moving the equipment to the outlet to obtain
the outlet data.
The optical particle counter was calibrated with polystyrene
latex spheres. The indicated diameters of the particulate in
the stack gas can differ from the true diameters because of
the effect of refractive index differences on results obtained
from the particle counter. In order to check the diameter
obtained for this effluent, the diffusion batteries were used
as sedimentation chambers, and particle diameters obtained
from calculated sedimentation rates were compared with the
indicated optical particle diameters. This comparison is shown
in Figure 3, using values for particle density of 1.0 and
2.0 grams/cm3 in the sedimentation calculations. The plant
personnel reported that particle densities are estimated to
range from about 1 to 4 grams/cm3, with 1.5 grams/cm3 an
estimated average for the outlet particulate. The comparison
indicates fair agreement between the sedimentation diameters,
which are independent of refractive index, and the equivalent
optical diameters. Figure 4 shows the optical and diffusional
sizing system. The sampling probe was heated to slightly
above the stack temperature at the outlet to avoid condensation.
Andersen impactors were used simultaneously at the precipi-
tator inlet and outlet on August 20, 21, 22, and 23.
Isokinetic sampling was performed at a single point for both
the inlet and outlet. Due to the extremely low mass loadings
at the outlet, it was necessary to operate the impactors
for approximately 16 hours in order to obtain weighable
quantities of particulate. Brink impactors were operated at the
8
-------
1.4
1.2
e
3.
(T
LU
H
UJ
1.0
0.8
g
1
UJ
0.6
a
UJ
CO
LINE INDICATING
PERFECT AGREEMENT
0.4
0.2
O SP GR = 2.0 IN SED CALC
A SP GR = 1.0 IN SED CALC
02 0.4 0.6 0.8 I.O
EQUIVALENT OPTICAL DIAMETER, ym
Figure 3. Comparison of sedimentation and
equivalent optical diameters
1.2
-------
Flowmeters
Cyclone Pump
Process
Exhaust
Line
Neutralizer
Flowmeter
Particulate
Sample Line
Cyclone
(Optional)
Manometer
Recirculated
Clean Dilution
Air
Filter
Dilution
\ Device
Diffusion
Battery
Orifice
Manometer
Aerosol
Photometer
Diffusional Dryer
(Optional)
Charge
Neutralizer Pressure
Balancing
Line
Pump
Bleed
Figure 4. Optical and diffusional sizing system
10
-------
inlet, but it was not practical to obtain data with this
device at the outlet. Since the gas phase contains conden-
sable hydrocarbons, gaseous fluorides, and water vapor, and
is near the water vapor saturation temperature, condensation,
evaporation, and chemical reaction pose potential interference
problems for impactor mass measurements. In an effort to
determine the order of magnitude of some of these potential
interferences, two Andersen impactor "blank" runs were made with
a filter prior to the impactors. The blank runs gave an esti-
mate of the weight loss or gain which could be expected due
to reactions between the gas phase and the fiberglass sub-
strates. Although the blank impactors were heated above the
stack temperature prior to sampling, condensation occurred in
the upper region of the impactor. The condensation was
apparently caused by relatively short-term temperature
variations in the outlet stack. For the runs used for size
determinations, the impactors were heated to about 120°F to
avoid the condensation problem.
Table 1 gives the weight changes obtained from the "blank"
impactor runs. No data were obtained with the first stage
blank due to the condensation problem. These blank changes
are not significantly greater than those which may normally
occur due to handling of the glass fiber substrates, and
are therefore not considered to pose a serious interference
problem.
MASS LOADING MEASUREMENTS
A modified EPA sampling train with an in-stack filter holder
(the same filter used for the EPA train) was used for the
mass loading measurements. The filter holder was teflon-
coated to avoid interference problems which might be caused
by corrosion of metal surfaces. Mass loading determinations
11
-------
Table 1. WEIGHT CHANGES OF ANDERSEN SUBSTRATE AFTER
SAMPLING FILTERED EFFLUENT FROM WET ESP
Stage Sampling Time
240 min 103 min
1 -
2 +0.06 mg +0.02 mg
3 -0.04 -0.04
4 -0.02 +0.04
5 +0.08 -0.04
6 -0.12 -0.16
7 -0.08 -0.10
8 -0.12 -0.10
Average, -0.04 mg
12
-------
were conducted at the inlet and outlet simultaneously with the
impactor runs. An isokinetic traverse across the stack was con-
ducted at both the precipitator inlet and outlet through a single
sampling port at each location for all but the last day of
the test series. On that date, a single point mass deter-
mination was performed at the outlet. As with the Andersen
impactors, it was necessary to heat the outlet filter holder
to approximately 120°F to avoid gross amounts of condensation.
However, the filters were still slightly damp (both inlet and
outlet), and consequently were placed in an oven at 120°F
for a few hours prior to desiccation and weighing.
13
-------
SECTION IV
RESULTS
IMPACTOR MEASUREMENTS
Tables 2 and 3 present results obtained from the Andersen
impactors during the four days of testing with these devices.
The outlet results are tabulated as the mass gain per stage
to enable comparison with the "blank" weight changes given
in Table 2. It should be noted that the weight changes for
the blanks are in general not proportional to the sampling
time. Although the blank changes represent a significant
fraction of the stage weights obtained during the outlet
sampling, there is sufficient mass to enable meaningful
conclusions to be drawn from the data. Figures 5 and 6 give
the mass loadings at the inlet and outlet respectively on a
cumulative basis, and Figure 7 gives the average inlet and
outlet size distributions from the Andersen impactor data on
log probability co-ordinates. No corrections were made for
the blank weight changes. The mass median diameters of both
inlet and outlet distributions are less than 1.0 urn. The
average outlet size distribution, and all subsequent calcu-
lations involving the outlet Andersen impactor measurements,
were obtained using runs 04, 05, and 06. Run 03 was discarded
because it appeared to collect an anomanously low amount of
mass when compared with the other three data sets.
In addition to the Andersen impactor measurements, several runs
were made at the inlet using a modified Brink impactor. Outlet
measurements with the Brink impactors were not practical due to
the long sampling times which are required due to the low mass
loading. Figure 8 gives a comparison of the averaged cumulative
loadings obtained with these impactors. Both impactors had
glass fiber substrates on the stages but the Andersens
were operated for approximately 20 minutes, whereas two hours
14
-------
Table 2. ANDERSEN INLET DATA
Run
No.
Date
Total Mass,
mg/amj
Lower Size
Limit, ym
10.0
7.01
4.33
3.05
1.99
0.93
0.56
0.40
AI-2
8/20/74
88.2
AI-4
8/20/74
60.7
AI-5
8/20/74
57.2
AI-7
8/21/74
39.7
AI-8
AI-9
8/21/74 8/21/74
73.8
83.2
AI-11
8/21/74
80.9
Cumulative Mass,
65.8
59.9
57.2
56.0
54.6
44.1
32.4
22.4
53.0
50.4
49.8
49.3
48.7
38.7
27.8
19.0
51.2
47.9
46.6
46.2
45.1
37.0
28.1
20.1
37.0
36.0
35.2
33.5
33.0
29.2
22.4
14.5
63.1
58.6
55.9
53.9
52.5
46.1
34.6
23.2
75.8
72.2
69.8
68.5
67.2
61.2
49.4
34.8
72.8
68.1
65.2
63.7
61.9
57.1
49.3
36.5
AI-12
8/22/74
128.5
mg/am 3
112.0
105.0
99.8
96.7
92.9
72.6
47.0
30.5
AI-14
8/22/74
82.9
74.9
72.5
71.8
71.5
70.3
62.3
48.3
31.0
AI-16
8/23/74
92.6
82.9
68.1
65.0
63.3
61.6
54.2
37.3
23.9
AI-17
Avg.
Avg.
8/23/74
100
91
85
81
78
74
60
37
23
.8
.1
.7
.1
.0
.0
.1
.8
.8
80.
70.
65.
63.
61.
60.
51.
37.
25.
77
87
85
4
87
16
14
67
43
87.7
81.52
78.49
76.60
74.48
63.32
46.64
31.48
-------
Table 3. ANDERSEN OUTLET DATA
Run Number
Stage
No.
1
2
3
4
5
6
7
8
F
Cut Point,
ym
10
7
4
3
2
1
0
0
.3
.2
.4
.1
.1
.0
.6
.4
03
mg
0.28
0.14
0.10
0.10
0.14
0.08
0.66
1.00
4.46
mg/am J
cum.
0.582
0.570
0.561
0.552
0.540
0.533
0.476
0.389
04
mg
0.34
0.26
0.20
0.26
0.28
0.46
1.26
2.18
4.78
mg/am J
cum.
0.847
0.824
0.807
0.748
0.759
0.719
0.609
0.418
05
mg
0.40
0.32
0.44
0.44
0.44
0.22
1.38
1.94
4.62
mg/am *
cum.
0.848
0.820
0.782
0.744
0.706
0.687
0.568
0.400
06
mg
0.44
0.34
0.44
0.32
0.40
0.40
1.10
1.58
4.18
mg/am3
cum.
0.719
0.691
0.655
0.628
0.596
0.563
0.472
0.343
Avg.
mg/am3
for
04,05,06
cum.
0.805
0.778
0.748
0.707
0.687
0.656
0.550
0.387
Total Mass Loading, mg/am:
0.606
0.877
0.882
0.755
0.838
-------
1000
T3
«J
O
10
U)
CO
4J
IQ
U
100
10
0.1
1.0 10.0
Diameter, ym
100.0
Figure 5. Inlet Andersen data
17
-------
-------
10
8
6
E
a.
£ 1.0
^-
^ 0.8
5 0.6
0.4
JNLET
OUTLET
0.2
0-1
12 S 10 20 30 40 50 80 9O 95 98 99
% SMALLER THAN INDICATED SIZE
Figure 7. Andersen data on log probability
co-ordinates
19
-------
100
CO
CO
w
w
(0
0)
•H
-U
0.1
1.0
Diameter, pm
10.0
Figure 8. Comparison of Andersen and Brink impactor
average inlet data
20
-------
were required for the Brink runs. The disagreement between
the two sets of measurements illustrated in Figure 8 is most
pronounced in the fine size range. Previous comparisons at
coal-fired power plants had shown reasonable agreement between
size distributions obtained at the same location with the two
impactors. Additional sampling on this particular effluent
and a stage-wise analysis of the collected material would
be required to explain the disagreement indicated by Figure 8.
Figures 9 and 10 are plots of dM/d log D from the Andersen
impactor measurements at the inlet and outlet, respectively.
Both of the distributions appear to be bimodal. The first
peak occurs at about the same particle diameter for both the
inlet and outlet data, but the second peak for the outlet is
shifted to the left on the diameter axis. These data were used
to obtain the efficiency as a function of particle diameter
given in Figure 11. The midpoints were obtained from the
average values of dM/d log D. The bands were obtained by: (1)
calculating the standard deviation at the indicated points for
the inlet and outlet data sets, (2) plotting dM/d log D values
which represent plus and minus one standard deviation from the
average at each particle diameter, (3) drawing curves through the
points representing plus and minus one standard deviation for
both inlet and outlet data sets, (4) calculate a minimum
efficiency for each diameter from
Minimum eff. = [(inlet average - la) - (outlet average + la)]100
inlet average - la J
(5) similarly, calculate a maximum efficiency
Haxi-nun, eff. = ; dnlet ""rl.) (outlet average - l.)]10<>
These maximum and minimum values are plotted as bars in Figure 11.
21
-------
1000
500
200
100
50
20
10
i i i i i i r
i i
0.1 0.2 0.5 1.0 2.0 5.0 10.0 20.0 50.0
Diameter, urn
Figure 9 . Inlet differential size distribution
22
-------
10.0
5.0
2.0
1.0
.5
.2
.1
o
rH
.05
02
01
1 I
1 I
I III
I I
0.1 0.2 0.5 1.0 2.0 5.0 10.0 20.0 50.0
Diameter, pm
Figure 10. Outlet differential size distribution
23
-------
K>
u
c
0)
o
•H
m
u
99.95
99.9
99.8
99.5
99
98
95
90
80
70
60
50
30
10
/
0.1 1.0 10 100
Diameter, \im
Figure 11. Andersen fractional efficiency data
1. Computed from modified field and diffusion charging calc.
2. Computed from a charging theory developed by Smith and McDonald3
3. Andersen data
-------
The apparent decrease in efficiency which occurs between 1.4
and 2.0 ym in diameter in Figure 11 is a reflection of the
second peak which occurs on Figure 10. Also plotted on Figure 11
are curves obtained from a mathematical model of an electro-
static precipitator developed by SRI under EPA contract. These
computer curves and the results obtained from the impactor
data are discussed further in a subsequent section.
It should be noted that the diameters reported here for the
inertial data are based on an assumed particle density of 2.0
grams/cm3. If the true densities are lower than this value,
the diameters as given should be increased by a factor equal
to the square root of the ratio of the assumed density to true
density.
OPTICAL AND DIFFUSIONAL MEASUREMENTS
Since it was necessary to obtain optical and diffusional data
at different times for the inlet and outlet, source stability
was investigated by obtaining particle concentration as a func-
tion of time data with the optical and diffusional sampling
system at the outlet. A representative data set is shown for
the condensation nuclei counter and optical particle counter in
Figures 12 and 13. The CN counter and the 0.3-0.5 ym channel on
the optical counter are reasonably stable, but the 0.5-0.7 ym
and the 0.7-1.3 ym channels show a considerable decrease with
time. However, the indicated variations are small in comparison
with those observed on effluents from other metallurgical
processes. These data suggest that the process was stable
enough to enable meaningful nonsimultaneous measurements.
Figure 14 gives the cumulative size distribution on a number
basis for this test series and several other sources which have
been tested by SRI with this equipment.
25
-------
NJ
C
o
•H
-P
1C
M
-P
C
0)
o
c
o
o
9.5
8.5
7.5
6.5
5.5
3 4.5
id
H
3.5
10
15 20 25
Time, Minutes
30
35
40
Figure 12
Relative Concentration Variation from Condensation
Nuclei Counter
-------
to
0)
•H
O
•H
4J
H
n)
04
4-1
o
100.000
10,000
1000
100
0.3-0.5 ym
0.5-0.7 ym
40
50
Figure 13,
10 20 30
Minutes
Relative concentration variation from
optical particle counter
27
-------
o
o
o.
O
Ul
o
o
o
Ul
3
o
108
107
6
I0
105
I03
102
OPEN HEARTH
FURNACE
0.01
Figure 14.
SUBMERGED
ARC FERRO-
' ALLOY FURNACE
INLET TO WET ESP
AT ALUMINUM
PLANT
\"S02 BUBBLE
\CAP SCRUBBER
PACKED BED
SCRUBBER
I
I
O.I 1.0 5.0
PARTICLE DIAMETER,urn
Cumulative size distributions on
a number basis for various
industrial particulate sources
as measured by optical and
diffusional methods
28
-------
JO
vo
Collection Efficiency, %
VO VO VO VO
vo vo vo vo
u> a\ vo vo vo vo • • • •
O O O VJl 00 VO Ol 00 VO VO
00
+
+
&
+
f\ & O
O u
1
A
0
MEASUREMEI*
ACASCAD
0 OPTICAL
4- DIFFUS
PRECIPITATO
TEMPERATU
C^ A CO
CURRENT D
A
£
JT METHOD.
E IMPACTORS
. PARTICLE COUN
IONAL
R CHARACTERIS1
RE— 4ICC
m2/( m3/sec)
ENSITY-30nA/(
0
A
TERS
'ICS-
im2
0.05
0.1
0.5 1.0
Particle Diameter, urn
5.0
10.0
Figure 15.
Measured fractional efficiencies for a wet electrostatic
precipitator with the operating parameters as indicated,
installed downstream of a spray type scrubber on an
aluminum reduction pot line.
-------
Fractional efficiencies were computed from the optical and
diffusional data, based on inlet measurements conducted on
August 20 and 21, and outlet measurements conducted on
August 22 and 23. Figure 15 gives the results of these
calculations, together with the inertially determined
fractional efficiencies. The optical and inertial effi-
ciency data show fair agreement over the size range 0.3 to
about 0.7 ym. However, when the optically determined
particle concentrations were converted to mass using a density
of 2.0 g/cm3, the mass loadings obtained were consistently
much lower than those obtained for the same size interval
with the Andersen impactors. These observations suggest
that the Andersen impactors were collecting mass which was
not detected by the optical particle counter at both the
inlet and outlet.
A pronounced increase in the collection efficiency is indi-
cated by the diffusional methods for particle sizes below
0.1 ym but larger than 0.05 ym. This behavior is consistent
with theoretical considerations and has been observed at
other installations utilizing electrostatic precipitators."
MASS LOADING MEASUREMENTS
Mass train measurements were obtained by Guardian Systems,
Inc., of Anniston, Alabama under subcontract to Southern
Research Institute on August 20, 21, 22, and 23. The results
of these measurements are given in Table 4. Results obtained
by a local pollution control agency on October 9-10, 1973
are given for comparative purposes in Table 5. In general,
fair agreement is expected between the total mass
loading obtained with cascade impactors and that obtained
with a mass train. A comparison of the total average mass
loading obtained with the Andersen impactors at the inlet
30
-------
Table 4. MASS TRAIN TEST RESULTS - 553 WET ESP
Ul
Run No.
Date
Sampling Time, mm.
H20, % by vol in gas
Avg. Gas Temp., °C
Flow, am3/sec
3Flow, DN m3/sec
mg/am3
gr/acf
gr/Dscf
Efficiency, %
Inlet
1
8/20
300
5.09
40.9
67.2
55.4
89.0
0.0389
0.0443
2
8/21
250
4.91
40.9
62.5
51.7
94.5
0.0413
0.0449
3
8/22
280
5.45
40.9
54.9
45.1
95.7
0.0418
0.0476
4
8/23
310
6.03
40.9
62.5
51.1
100.9
0.0441
0.0494
1
8/20
376
5.19
234.2
44.6
37.6
4.58
0.00200
0.00220
95.03
Outlet
2
8/21
375
4.96
38.1
43.5
36.3
4.26
0.00186
0.00209
95.34
3
8/22
375
5.22
38.1
44.4
36.9
3.57
0.00156
0.00166
96.51
4"
8/23
360
5.95
38.1
43.9
36.3
1.97
0.00086
0.00098
98.02
Notes:
1.
2.
3.
4.
Based on traverse across one sampling port and area of 3.05 mz(32.85 ft2) - see text.
Based on traverse across one sampling port and area of 3.54 mz(38.10 ft2).
08C and 760 mm Hg.
Obtained at a single point near the center of the stack.
-------
(Table 2) with the average inlet mass loading from Table 4
indicates that the impactors collected about 90% of the
material collected by the mass train. A total average mass
loading of 0.0426 gr/acf was reported by Hofer5 from 37 tests
on the outlet of one of the spray towers at this plant site.
These results are consistent with those reported in Tables
2 and 4. The inlet gas flows reported in Table 4, however, are
anomalously high. Possible reasons for these results are:
1) there may have been an undetected calibration error in the
stack sampling system used at the inlet, 2) the velocity profile
obtained at the single port available for the mass train
measurements may be nonrepresentative of the average flow.
The outlet flow rates reported in Table 4 are considered to
be the correct flow rates since they show good agreement
(within about 5%) with those obtained by a local pollution
control agency.
In contrast to the agreement shown between mass loadings
obtained with the Andersen impactors and the mass train at
the inlet, severe disagreement was obtained at the outlet.
The total mass obtained with a traverse using the mass train
at the outlet was greater than that collected with the
impactors by a ratio of approximately 5 to 1. When the mass
filter was operated near the center of the stack and the
sampling location used for the impactors, the disagreement
was reduced to a ratio of about 3 to 1. A comparison of
outlet loadings between Tables 4 and 5, however, indicates
that the mass train results obtained during this test series
are in fair agreement with those obtained previously by a
local pollution control agency. Note that the Andersen data
in Table 3 and the mass data in Table 4 show good repro-
ducibility.
32
-------
Table 5. RESULTS FROM TESTS CONDUCTED BY A LOCAL POLLUTION
CONTROL AGENCY
October 9 and 10, 1973
'Total particulate, LVS
2Total particulate, IVS
Percent water vapor
Gas flow
0.0029 gr/oscf =7.02 mg/DNm3
0.00208 gr/Dscf=5.03 mg/DNm3
5.2%
99,000 acfm = 46.7 m3/sec
!Low volume sampler.
2Intermediate volume sampler.
33
-------
Reasons which have been hypothesized for the disagreement
between the Andersen impactor and the mass train data are:
• The conditions in the impactor lead to evapora-
tion of gross amounts of previously condensed
hydrocarbons.
• Relatively large water droplets, containing about
5% by weight of dissolved solids, were collected
by the mass filter, but not by the impactor.
Evaporation of these droplets would leave a
residue which could account for the greater mass
observed with the mass filter.
In an effort to resolve the disagreement, the substrates from
one Andersen run and the outlet filters from runs 3 and 4 of
Table 4 were submitted to Southern Research Institute's
Analytical Services Section for analysis with a gas chromato-
graph (GC). The objective of this analysis was to determine
the relative volatility and approximate mass, if possible, of
the hydrocarbons remaining on the filters and fiberglass
substrates. The instrument and conditions used were as
follows:
Instrument: Hewlett Packard Model 5750B,
with Flame lonization Detector
Column: 10% UC W-98 on Gas Chrom Q
Temperature: 25°C to 250°C at 30°C/min-held at
maximum
Standards: Cg, C12, C16' C22 in CS2 at con~
centration of 100 yg/ml CS2
Sensitivity: ^1 yg/ml sample
Extraction: Extract filter or substrate, or a
portion thereof, with 1 ml of
CS2
34
-------
Five ml of the CS2 extract were injected into the GC at the
conditions stated above. Analysis of the standard solutions
of hydrocarbons under these same conditions gave the follow-
ing results.
Compound Retention Time, min
Octane, Ca 4.3
Dodecane, Ci2 7-4
Octadecane, GIS 11.4
Docosane, C22 20.0
The retention time of extracts of the filters and substrates
indicated that very little of the hydrocarbons were in the
Cg to Ci2 retention time range. The major components were
eluted at times greater than that indicated for Cie- It is
apparent from these results that the hydrocarbons remaining
on both the filters and substrates are relatively non-volatile,
and therefore, the discrepancy cannot be explained by com-
paring the volatility and mass of the hydrocarbons remaining.
It is possible, however, that if the above analyses were
conducted immediately upon removal of the sampling devices
from the stack, that significant differences may have been
observed between the hydrocarbons on the filters and substrates,
It is our conclusion that the most probable cause of the mass
loading discrepancy is the collection of large water droplets
containing solids by the mass filter. Such droplets would be
subject to stratification in the stack, and this is quali-
tatively indicated by the decrease in loading which occurred
when the mass train was operated at a single point. Addi-
tional work with a traverse using a sampling device designed
to provide sizing information above 10 pm diameter would be
required to resolve the problem.
35
-------
SECTION V
DISCUSSION
DESCRIPTION OF THE WET ELECTROSTATIC PRECIPITATOR
The wet electrostatic precipitator on which this
test series was conducted is a wire and plate design with
three electrical sections in series in the direction of gas
flow. Plate-to-plate spacing is 30.5 cm (1 ft), and each
collecting electrode is 1.83 m long (6 ft) and 7.52 m high
(25 ft). Thus, the total parallel plate collecting elec-
trode length is 5.48 m, or 18 ft. Each electrical set
powers 28 gas passages. Figures 16 and 17, taken from the
manufacturer's literature, illustrate the overall precipi-
tator arrangement and the electrode design, respectively.
The total parallel plate collecting area is 2342 m2(25,200
ft2), and the "transverse baffles", which are perpendicular
to the gas flow, provide additional collecting electrode
area. The effective collecting area provided by these
baffles was estimated as 390 m2(4200 ft2), resulting in a
total collection area of 2732 m2(29,400 ft2). Average
specific collecting area during the test series was there-
fore 62 m2/(m3/sec), or 315 ft2/(1000 cfm) .
Electrode irrigation is provided by sprays at the precipi-
tator inlet and above the collection plates. The sprays
provide a mist which is collected along with the particu-
lates in the flue gas, and the electrode cleaning is accom-
plished by the coalescence and subsequent downward flow of
the collected spray droplets. The sprays are operated
continuously, except for those installed near the precipitator
outlet, which are operated only periodically. These spray
nozzles were not in operation during the test program.
36
-------
UJ
Figure 16. Wet electrostatic precipitator
-------
Transverse Baffles
B Discharge Electrode
3'fC Collecting Electrode
D Discharge Electrode
E Transverse Baffles
Figure 17. Schematic of electrode arrangement
38
-------
Table 6, provided by the manufacturer,1 summarizes the
specifications for the wet precipitator installation. The
irrigating fluid is a high pH sodium'based liquid which is
returned to clarifiers and a cryolite recovery plant.
Plant personnel reported that the cryolite recovery
system is essentially a closed liquid loop, which results
in a solids content of about 5% by weight being returned
to the wet ESP - scrubber system. Liquor flow through the
wet ESP during the test program was constant at 31.5 I/sec
(500 gal/min), which gives a liquid to gas ratio of about
0.7 1/m3 (5.3 gal/1000 ft3). Liquor temperature, based on
measurements reported by plant personnel, ranges from
90 to 104°F, and is usually 94 to 95°F. No significant
temperature drop has been observed in the liquor loop
across the precipitator.
ELECTRICAL CONDITIONS
Voltage and current readings were obtained from the panel
meters of the 553 precipitator periodically during the test
program. These data are given in the Appendix, Table A-2.
At the conclusion of the test program, voltage-current
curves were obtained for the unit with the spray system
operating normally. The secondary voltage-current relation-
ships are given in Figure 18, along with the range of opera-
tion that was observed for each electrical set during the
test program. The difference between the voltage-current
curves and the operating ranges is a result of the fact
that, in normal operation, the power supplies are operating
under automatic control with a certain spark rate, whereas
the V-I curves were obtained by manually increasing the
applied voltage until sparking occurred. The plant personnel
were operating the power supplies at a spark rate which was
believed to maximize the time-averaged electric field.
39
-------
Table 6. SUMMARY OF SPECIFICATIONS FOR THE WET
ELECTROSTATIC PRECIPITATORS l
Gas Flow
Inlet Temperature to Scrubbers
Inlet Temperature to WESP
Total Particulate Inlet Loading (solids
and condensables, excluding water)
No. of Electrostatic Fields
Liquor, Flow Rate at 60 psi (5.08 atm.
absolute)
Liquor, pH in
Outlet Loading for an Inlet Loading of
0.05 gr/scf or less (0.114 g/m3)
Minimum Collection Efficiency for Outlet
Loadings Greater than 0.003 gr/scf
(0.0069 g/m3)
Face Velocity
Maximum Pressure Drop
Treatment Time
100,000 scfm, or
47.2 m3/sec at
standard conditions
121°C
38.1 - 43.7°C
0.05 gr/scf, or 0.114
g/m3 at standard conditions
500 gpm or 31.5 I/sec
7-10
0.003 gr/scf, or 0.0069
g/m3 at standard conditions
95%
2.38 ft/sec, or 0.726 m/sec
1" W.G., or 2.54 cm
10.1 sec
Housing Material, Hot Rolled MS, Thickness 3/16", or .476 cm
10 GAUGE
Collection Plates, Hot Rolled MS,
Thickness
Discharge Electrodes, Flatbars MS
Piping Materials
Spray Nozzles, SS 316, Type
No. of Transformer Rectifiers
Rectifier Type
Wave Form
Minimum Output per T-R Set
Primary Voltage
1" x 1/8", or 2.54 cm x
0.318 cm
PVC
Full Cone
3
Silicone
Full
60 kV, 1000 ma
480 V, 60 Hz
Manual and Automatic Voltage and Spark Rate Control
40
-------
20 24 32 36 40 44 48 52 56
Figure 18. Voltage current relationship (Manual Control)
and operating ranges (Automatic Control)
41
-------
The V-I curve for the first electrical set is shifted toward
high voltages for a given current when compared with readings
from the other electrical sets. This behavior is often
observed and is a reflection of the higher space charge
density contributed by the higher particulate loadings
which exist in the inlet field. Although the third field
operates at a relatively high current, the average current
density for all three sets was only about 30 na/cm2. The
current density limitation was imposed by sparking, since
the electrical resistivity of the particulate is not a
factor in the wet mode of operation.
COMPARISON OF RESULTS WITH THEORETICAL PREDICTIONS
Figure 11 presented the inertially determined fractional
efficiencies and two predicted curves obtained from a theo-
retically based computer model of an electrostatic precipi-
tator.6 This mathematical model calculates theoretically
expected collection efficiencies for representative particle
diameters as a function of precipitator operating conditions.
Predicted collection efficiencies for each particle diameter
are a function of the electric field, the charge on the
particle, and the ratio of collection area to gas volume
flow rate. Curve 1 was based on a procedure for calcu-
lating particle charge as described by Gooch and Francis,6
whereas curve 2 was based on a recently developed particle
charging theory described by Smith and McDonald.3 The
latter charging theory gives better agreement with the
available experimental data on particle charge over the
indicated diameter range, and is therefore considered to be
the preferred basis for predicting collection efficiencies.
It can be seen that fair agreement is obtained between curve
2 and the inertially determined efficiencies over the particle
diameter range 0.25-1.3 vim, but that the measured values depart
42
-------
drastically from the predictions at diameters larger than 1.5 ym.
This apparent departure from the expected functional form may be
caused by the generation of particles within the device, possibly
originating from the liquid sprays or from reentrained liquid
that is not captured by the outlet transverse baffles, which
are considered by the manufacturer to function as an electro-
statically augmented mist eliminator. It should be noted that
the diameter band 0.25-1.3 um, based on the Andersen measure-
ments, represents 54% of the mass at the inlet and 56% of the
mass at the outlet.
Since a major portion of the particulate entering the precipi-
tator is known to consist of condensed hydrocarbons, it is of
interest to consider the effect of dielectric constant on
predicted collection efficiencies. The predictions shown in
Figure 11 were based on the assumption that the particulate in
the wet environment may be characterized by high values of di-
electric constant. In order to examine the effect of low values
of dielectric constant on the predicted efficiencies, the
computer program for calculating particle charge used in
obtaining curve 2 on Figure 11 was employed with dielectric
constants (E) of 2 (the lowest value which might be representa-
tive of a hydrocarbon droplet) and 100. The results of -these
calculations are presented in Table 7.
Table 7. EFFECT OF DIELECTRIC CONSTANT
ON PREDICTED PENETRATION
(Smith-MacDonald Theory Used for Calculating
Particle Charge)
Penetration, Penetration,
Particle Diameter, % %
ym for e = 100 for e = 2
0.2 2.95 3.45
0.50 1.135 1.88
0.70 0.384 0.82
1.30 0.011 0.05
1.70 9.1x10-** 0.007
43
-------
It can be seen that this range of variation of dielectric
constant has a significant effect on predicted performance,
with the largest effect being observed for the larger
particles. Since the particulate consist of both organic
and inorganic matter in a wet atmosphere, it is reasonable
to expect a major portion of the mass would exhibit a
relatively high dielectric constant under these conditions.
Electrostatic precipitator performance is often described
by an empirical performance parameter termed the precipi-
tation rate parameter. The parameter is obtained by evaluat
ing the Deutsch equation using the overall mass efficiency
and the ratio of volume flow to plate area:
Evaluation of this relationship using the data in Table 4
gives the results presented in Table 8. A predicted precipi-
tation rate parameter may be obtained from the computer model
based on the inlet size distribution obtained from the
Andersen impactor measurements. Based on the predicted
efficiencies indicated by curve 2 of Figure 11, numerical
integration over the inlet size distribution gives a total
predicted penetration of 1.1% (98.9% efficiency), and
predicted precipitation rate parameter of 7.3 cm/sec, which
shows fair agreement with the data in Table 8. Figure 11
shows, however, that the model underpredicts fine particle
collection efficiencies, and overpredicts collection for
particles larger than about 0.60 pm.
44
-------
Table 8. PRECIPITATION RATE PARAMETERS
Run No.
Gas flow, m3/sec
Mass
Efficiency,
Precipitation
Rate
Parameter,
cm/sec
1
2
3
4
44.6
43.5
44.4
43.9
95.03
95.34
96.51
98.02
4.90
4.88
5.45
6.30
QUALITATIVE COMPARISON OF MASS MEASUREMENTS
Results obtained from performance tests on precipitators
collecting fly ash from coal-fired boilers usually indicate
that: 1) the mass obtained from the particle size measuring
techniques used during this test series are in fair agree-
ment with each other in the regions of overlap, and 2) the
total mass from the impactor measurements is in fair agree-
ment with the mass train determinations. The disagreement
found during this test series is summarized qualitatively
in Table 9.
Table 9. QUALITATIVE COMPARISON OF MASS MEASUREMENTS
Inlet
Outlet
Mass train > Andersen
Andersen = mass train
Brink < Andersen
Andersen > optical Andersen > optical
Andersen efficiency = optical efficiency
45
-------
Since the results obtained by each measurement technique were
fairly reproducible, we conclude that the disagreement was
caused by sampling procedures or conditions peculiar to each
of the instruments. The Andersen impactor results appear to
provide the best data for fine particle collection efficiencies
and mass loading in the fine particle range, whereas the mass
train data should be used to obtain the overall particulate
collection efficiency and the total outlet mass loading.
The optical data are believed to be less reliable, since
dilution and out-of-stack sampling are required to obtain
results.
COST ESTIMATES
The estimates operating power required for operation of the
wet electrostatic precipitator is given in Table 10. If
power costs are $0.01/kwh, the power costs would be about
$27.00 per day of operation for the precipitator. Bakke l
has reported that the installed flange to flange capital
costs of the wet precipitator are between $3.00 and $4.00
per cfm, based on mild steel construction. The operators
reported that their total costs for installing the wet
precipitators at the reduction plant would approximate
$18,000,000 or about $6.00/cfm.
46
-------
Table 10. OPERATING POWER ESTIMATED FOR WET
ELECTROSTATIC PRECIPITATOR NO. 553
Item Basis Power, kw
Power supplies Primary meter readings 49.0
Pumping power 100 psig total head, 31.5
I/sec, 60% pump efficiency 36.0
Fan power 1.27 cm H2O AP, 50% fan
efficiency, 44.1 m3/sec 11.0
Insulator heater
power 6 kw/field, from Bakkel 18.0
TOTAL 114.0 kw
47
-------
SECTION VI
REFERENCES
1. Bakke, Even. The Application of Wet Electrostatic
Precipitators for Control of Fine Particulate Matter.
Paper presented at the Symposium on Control of Fine
Particulate Emissions from Industrial Sources for
the joint U.S.-U.S.S.R. Working Group, Stationary
Source Air Pollution Control Technology.
San Francisco, California, January 15-18, 1974.
2. Smith, W. B., K. M. Gushing, and J. D. McCain.
Particulate Sizing Techniques for Control Device
Evaluation, Special Summary Report. Southern
Research Institute report to the Environmental
Protection Agency under Contract No. 68-02-0273,
July 12, 1974.
3. Smith, W. B. and J. R. McDonald. Calculation of the
Charging Rate of Fine Particles by Unipolar Ions .
Journal of Air Pollution Control Association. 25(2)
February 1975.
4. McCain, J. D., J. P. Gooch, and W. B. Smith. Results
of Field Measurements of Industrial Particulate Sources
and Electrostatic Precipitator Performance. Journal of
Air Pollution Control Association. 25(2), February 1975
5. Hofer, George Charles. Relationship of Operating
Parameters to the Efficiency of a Centrifugal Spray
Tower for the Collection of Particulates Emitted from
a Horizontal Spike Soderberg Aluminum Plant. Master
of Science Thesis in Civil Engineering, University of
Washington, 1971
48
-------
6. Gooch, J. P. and N. L. Francis. A Theoretically-Based
Mathematical Model for Calculation of Electrostatic
Precipitator Performance. Journal of Air Pollution
Control Association. 25(2), February 1975.
49
-------
SECTION VII
APPENDIX
50
-------
TABLE A-l
POT OPERATIONS DURING TESTS
Date
Operation
Pot No.
Monday, 8/19/74
Tuesday, 8/20/74
Wednesday, 8/21/74
Thursday, 8/22/74
Friday, 8/23/74
Pin and channel pull 630,633,639,553
Flex raise
632,555
Flex raise 638
Pin and channel pull 632,637,555
Flex raise
Pin and channel pull
Flex raise
Pin and channel pull
629,640,641
638
543
629,640,641
51
-------
TABLE A-2
WESP #553
Voltage-Current Readings
Field No. 1
Prim. Amp.
35-45
50
40-50
40-50
40-50
40-55
40-55
30-55
35-55
40-55
40-55
40-55
50-60
55-65
35-55
35-55
40-50
Prim. Volts Sec mA
290-310
300-310
280-320
280-300
260-300
280-310
280-300
280-320
280-320
260-300
280-320
260-300
250-280
260-300
290-330
280-320
260-320
100-250
200-250
^200
150-200
200-300
150-250
200-300
150-250
150-250
200-250
150-250
150-220
180-250
200-250
150-250
150-250
150-250
Sec kV
40-48
42-48
46-50
42-48
42-48
46-50
44-48
42-50
42-50
44-48
42-48
44-48
42-48
42-48
40-48
42-48
44-48
Date Time
8/19/74 11:
1:
8/20/74 9:
11:
2:
" 4 :
6:
8/21/74 8:
11:
2:
5:
8/22/74 8:
1:
3:
8/23/74 8:
1:
4:
30 am
30 pm
15 am
55 am
25 pm
25 pm
25 pm
50 am
10 am
50 pm
30 pm
45 am
55 pm
40 pm
30 am
10 pm
20 pm
52
-------
TABLE A-2
(Continued)
WESP #553
Voltage-Current Readings
Field No. 2
Prim. Amp.
40-50
55
45-55
55-65
50-60
30-40
40-50
30-45
30-40
40-55
45-55
50-65
50-60
55-65
45-55
50-65
40-55
Prim. Volts
240-280
260-280
260-280
240-280
240-280
210-230
200-220
210-230
200-220
240-280
220-260
270-300
250-280
260-300
240-280
260-300
260-300
Sec mA
150-240
200
150-200
200-250
150-250
100-150
100-150
100-150
100-180
180-250
180-250
200-280
180-250
200-250
200-280
200-260
180-240
Sec kV
40-45
40-44
40-46
40-48
40-48
40-44
40-44
40-44
40-44
42-48
42-48
42-48
42-48
42-48
40-46
40-48
42-46
Date
8/19/74
ll
8/20/74
ii
H
ii
n
8/21/74
II
11
II
8/22/74
n
n
8/23/74
ii
n
Time
11: 30 am
1: 30 pm
9:15 am
11:55 am
2:25 pm
4 : 25 pm
6:25 pm
8:50 am
11:10 am
2:50 pm
5:30 pm
8:45 am
1:55 pm
3:40 pm
8:30 am
1:10 pm
4:20 pm
53
-------
TABLE A-2
(Continued)
WESP #553
Voltage-Current Readings
Field No. 3
Prim. Amp.
90-100
90-100
90-110
90-110
90-105
90-100
90-110
90-110
80-100
95-105
95-105
90-110
95-110
100-110
90-110
95-110
95-110
Prim. Volts
320-360
340
320-340
320-360
340-360
320-360
310-330
340-360
300-360
340-380
320-360
300-340
320-360
340-360
340-380
340-360
330-360
Sec mA
350-600
500
400-550
450-550
400-500
400-550
400-550
450-550
400-500
400-500
480-550
350-500
450-550
450-550
400-550
450-550
450-550
Sec kV
50-54
50-54
50-54
50-54
50-56
50-56
50-54
48-54
50-58
50-56
50-56
48-56
48-56
48-56
50-58
50-56
50-56
Date
8/19/74
it
8/20/74
H
H
n
H
8/21/74
ti
n
11
8/22/74
n
M
8/23/74
n
n
Time
11:30 am
1:30 pm
9:15 am
11:55 am
2:25 pm
4:25 pm
6 : 25 pm
8:50 am
11:10 am
2:50 pm
5:30 pm
8:45 am
1:55 pm
3:40 pm
8:30 am
1:10 pm
4:20 pm
54
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-650/2-75-033
4 TITLE AND SUBTITLE
Particulate Collection Efficiency Measurements on a
Wet Electrostatic Precipitator
3 RECIPIENT'S ACCESSION-NO.
5 REPORT DATE
March 1975
6 PERFORMING ORGANIZATION CODE
7 AuTHORisi j p Qooch and J.D. McCain
Southern Research Institute, Birmingham, AL 35205
8 PERFORMING ORGANIZATION REPORT NO
SORI-EAS-74-415
3296-1
9 PERFORMING ORG'VNIZATION NAME AND ADDRESS
The M.W. Kellogg Co.
1300 Three Greenway Plaza
Houston, Texas 77046
1O PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-004
11 CONTRACT/GRANT NO
68-02-1308, Task 21
12 SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13 TVPE OF REPORT ANO PERIOD COVERED
Final Task; 7-12/74
14 SPONSORING AGENCY CODE
15 SUPPLEMENTARY NOTES
16 ABSTRACT*]^ report gives results of fractional and overall particulate collection effi-
ciency measurements of a plate-type wet electrostatic precipitator (ESP) collecting
fume from an aluminum pot line. Overall collection efficiencies, based on a mass
train with an in-stack filter,ranged from 95. 0 to 98. 0%. The mass filter obtained
much higher total outlet mass loadings than did the Andersen impactors, presumably
because of large entrained liquor droplets which were captured by the mass traverse,
but not by the single-point impactor measurements. The average minimum collection
efficiency in the size range 0.2 to 1.0 micrometer diameter (based on the Andersen
data) was 98. 5%. Comparisons between measured (with Andersen impactors) and
predicted collection efficiencies obtained from a mathematical model of an ESP indi-
cated fair agreement in the size range 0. 2 to about 1. 3 micrometers. For larger
particles, the collection-efficiency/particle-size relationship departed drastically
from the expected pattern, possibly because of liquor carryover from the electrode
irrigation system.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b IDENTIFIERS/OPEN ENDED TERMS
c. COSATI field/Group
Air Pollution
Dust
Measurement
Electrostatic Precipitators
Collection
Efficiency
Air Pollution Control
Stationary Sources
Particulate
13B
11G
14B
S DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unclassified
Unlimited
21 NO OF PAGES
60
2O SECURITY CLASS (Thispage)
Unclassified
22 PRICE
EPA Form 222O-1 (9-73)
55
------- |