EPA-600/2-78-062
March 1978
Environmental Protection Technology Series
EVALUATION OF FOUR
NOVEL FINE PARTICULATE
COLLECTION DEVICES
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental 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.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-78-062
March 1978
EVALUATION OF FOUR
NOVEL FINE PARTICULATE
COLLECTION DEVICES
by
S. Calvert, S. C. Yung, H. Barbarika,
and R. G. Patterson
Air Pollution Technology, Inc.
4901 Morena Boulevard, Suite 402
San Diego, California 92117
Contract No. 68-02-1496
ROAP 21ADL-004
Program Element No. 1AB012
EPA Project Officer Dale L. Harmon
Industrial Environmental Research Laboratory
Office of Energy, Minerals and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
ABSTRACT
Experimental performance evaluations were conducted on four
novel fine particulate control devices. They were the Johns-
Manville Cleanable High Efficiency Air Filtration System (CHEAP),
the APS Electrostatic Scrubber, the APS Electrotube, and the
TRW Charged Droplet Scrubber.
The performance evaluations included the measurements of
inlet and outlet particle size distribution and concentration with
cascade impactor and diffusion battery. Fine particle collection
efficiencies as functions of particle size were computed from the
data. Mathematical performance models were developed for the
CHEAP and the Electrostatic Scrubber. The models gave satisfac-
tory predictions.
An experiment was carried out in the laboratory to determine
the effects of charged particles on cascade impactor data. Re-
sults indicated that using the value of impaction parameter deter-
mined for uncharged particles will cause overesrimation of the
size of the charged particles collected.
111
-------
CONTENTS
Abstract 1:L1
Figures v
Tables viii
Abbreviations and Symbols x
Acknowledgement X1V
Sections
1. Introduction 1
2. Summary, Conclusions and Recommendations 2
3. CHEAP 6
4. APS Electrostatic Scrubber 20
5. APS Electro-Tube 37
6. TRW Charged Droplet Scrubber 47
7. Performance Test Method 59
8. Data Reduction and Computation 69
9. Effects of Charged Particles on the Experimental
Performance of Electrostatic Devices
87
References ......................... y
Appendix
IV
-------
FIGURES
No.
1 Schematic of CHEAP Unit 7
2 Inlet size distributions 10
3 Inlet and outlet size distribution 10
4 Run 18 wet and dry size distributions 11
5 Run 19 wet and dry size distributions 11
6 Penetration versus dry particle diameter
for run 10 13
7 Penetration versus dry particle diameter
for run 11 13
8 Penetration versus dry particle diameter
for run 12 14
9 Experimental grade penetration curve for CHEAP . . 16
10 Cut diameter versus pressure drop for fibrous bed . 18
11 APS Electrostatic Scrubber 21
12 Penetration versus particle diameter for data
set A 27
13 Penetration versus particle diameter for data
set B 27
14 Penetration versus particle diamter for data
set C 28
15 Penetration versus particle diameter for data
set A, B, 5 C 28
16 Penetration versus particle diameter using both
diffusion battery and cascade impactor, run 27
(set C, ionizer off) 29
17 Penetration versus particle diameter using both
diffusion battery and cascade impactor, run 28
(set C, ionizer off) 29
18 Penetration versus particle diameter using both
diffusion battery anc cascade impactor, run 29
(set C, ionizer on) 30
v
-------
FIGURES (continued)
M« Page
No. —£—
19 Penetration versus particle diameter suing both
diffusion battery and cascade impactor, run 30 30
(set C, ionizer on)
20 Efficiency of single drop versus inertial parameter
NR , = 9.6 with NpD as parameter .........
21 Efficiency of a single drop, n, versus NpD with
K as a parameter (NRed =9.6) ..........
22 Experimental and predicted particle penetration
versus particle diameter .............
T O
23 Diagram of APS Electro-Tube ............ -30
24 Penetration versus aerodynamic particle diameter
for low gas flow, runs 15, 16, 19 .........
25 Penetration versus aerodynamic particle diameter 45
for medium gas flow, runs 7, 10, 12, 13, 14 ....
26 Penetration versus aerodynamic particle diameter ^^
for high gas flow, runs 3, 4, 5, 18 ........
48
27 TRW Charged Droplet Scrubber schematic ......
28 Particle penetration versus aerodynamic diameter 5&
for low electrode voltage and low gas flow rate . .
29 Particle penetration versus aerodynamic diameter
for high electrode voltage and high gas flow rate .
30 Particle penetration versus aerodynamic diameter ^
for high electrode voltage and low gas flow rate. .
31 Particle penetration versus aerodynamic diameter 57
for low electrode voltage and high gas flow rate. .
' Modified EPA sampling train with in-stack cascade 62
impactor .....................
64
-5 Schematic diagram of diffusion battery system . . .
Particle diameter versus particle aerodynamic j-,
diameter .....................
Overall penetration fraction versus "^-./d N" 81
with "a " as a parameter ....... ?.?....
o
VI
-------
FIGURES (continued)
No.
36
37
38
39
40
41
A-l
A- 2
A- 3
A- 4
A- 5
A- 6
A- 7
A-8
A- 9
A-10
A-ll
A-12
A-13
A-14
Particle penetration through SDB at Q= 4.7 £pm . .
A.P.T. cut/power plot
Overall penetration versus cut to mass mean
particle ratio for log-normally distributed
particles. (Calvert et al. (1972)
Impaction characteristics with glass fiber
filter (0.5 ym PSL) . .
Size distribution of particles exiting ESP when
d^,, = 10 ymA, a =4.0
PgIN «
Fractional collection efficiencies for a full-scale
precipitator on a coal-fired power boiler. (Gooch
et al. (1975))
Inlet and outlet size distribution for run 1 ...
Inlet and outlet size distribution for run 4 ...
Inlet and outlet size distribution for run 6 ...
Inlet and outlet size distribution for run 8 ...
Inlet and outlet size distribution for run 9 ...
Inlet and outlet size distribution for run 10. . .
Inlet and outlet size distribution for run 13. . .
Inlet and outlet size distribution for run 14. . .
Inlet and outlet size distribution for run 15. . .
Inlet and outlet size distribution for run 16. ..
Inlet and outlet size distribution for run 17. . .
Inlet and outlet size distribution for run 18. ..
Inlet and outlet size distribution for run 19. . .
Diffusion battery data for inlet run 1 § 4 ....
Page
82
84
86
88
90
91
99
99
100
100
101
101
102
102
103
103
104
104
105
105
VII
-------
TABLES
1 Data Set A. Runs 1-10 ............... 25
2 Data Sets B, C, and D, Runs 11-31 ......... 25
3 Inlet and Outlet Size Distribution Summary .... 26
4 Test Run Summary ................. 26
5 Operating Conditions ............... 40
6 Size Distribution, Mass Loading, and Overall
Penetration Data ................. 42
7 Total Filter Particle Loading Tests
42
8 Number Basis Size distribution Data for Diffusion
Battery Tests ................... 43
9 Operating Conditions ............... ->0
10 Test Run Summary ................. 51
11 Cross-Reference to TRW Test Matrix ........ 52
12 Inlet and Outlet Size Distribution Summary . . . . 54
A-l Inlet and Outlet Sample Particle Data for Run 1 . . 95
A- 2 Inlet and Outlet Sample Particle Data for Run 4 . . 95
A- 3 Inlet and Outlet Sample Particle Data for Run 6 . . 95
A-4 Inlet and Outlet Sample Particle Data for Run 8 . . 95
A- 5 Inlet and Outlet Sample Particle Data for Run 9 . . 96
A-6 Inlet and Outlet Sample Particle Data for Run 10 . 96
A- 7 Inlet and Outlet Sample Particle Data for Run 13 . 96
A-8 Inlet and Outlet Sample Particle Data for Run 14 . 96
A- 9 Inlet and Outlet Sample Particle Data for Run 15 . 97
A- 10 Inlet and Outlet Sample Particle Data for Run 16 . 97
A-ll Inlet and Outlet Sample Particle Data for Run 17 . 97
Vlll
-------
TABLES (continued)
No. Page
A-12 Inlet and Outlet Sample Particle Data for Run 18. . 97
A-13 Inlet and Outlet Sample Particle Data for Run 19. . 98
IX
-------
LIST OF ABBREVIATIONS AND SYMBOLS
Latin
a - Constants, dimensionless
n
A - Constant
b - Weibull slope
B - Constant
C' - Cunningham correction factor, dimensionless
C,, - Drag coefficient, dimensionless
CDF - Cumulative distribution function, dimensionless
= 1 - exp
3
C^ - Mass of particles, g
C - Total mass of particles, g
CT - Total mass concentration or loading, g/cm
d - Symbol for differentiation
d - Wire diameter, cm
d C
£ - Slope of the cumulative mass versus particle diameter
cf(d ) curve at d , dimensionless
dj - Drop diameter, cm
df - Fiber diameter, cm
d- - Jet diameter, cm
d - Actual particle diameter, cm or ym
d - Aerodynamic particle diameter, ymA
pa
d - Aerodynamic cut diameter, ymA
pac
-------
ABBREVIATIONS AND SYMBOLS (continued)
d - Cut diameter, ym
d - Mass median diameter, ymA
r o
d - - Specific particle size, ym
d N - Number median particle diameter, ym
d - Minimum particle diameter, ym
E - Field strength, kV/cm
EQ - Charging electric field strength, kV/cm
f - An empirical constant, dimensionless
fa - Initial value of f - 0.5, dimensionless
K - Impaction parameter, dimensionless
I - Fiber pad thickness, cm
mp - Mass of particles in the infinitesimal size range
M . - Cumulative mass concentration of particles smaller
pl than dpi, g/cm3
N - Total number of particles, dimensionless
n - number of particles, dimensionless
Npp - Flux deposition number =
UF
—, dimensionless
o
N - Cumulative number concentration of particles smaller
v than d , no./cm3
N t - Total number concentration of particles, no./cm3
Pt(d ) - Penetration as a function of particle size, fraction
Ft - Overall particle penetration, fraction
Pt^ - Particle penetration for particle size "i", fraction
W
P(x) = rr- = Cumulative mass fraction of sizes smaller than
mt "d "
pa
xi
-------
ABBREVIATIONS AND SYMBOLS (continued)
- Vnl
G
Q - Volumetric flow rate of gas, m3/min
Q - Volumetric flow rate of liquid, mVmin
LJ
Q - Electrical charge carried by the particle, coulomb
S - Solidity, dimensionless
UT-. - Particle drift velocity, cm/s
r
Up - Superficial gas velocity, cm/s
Up. - Venturi throat gas velocity, cm/s
u- - Jet velocity, cm/s
u - Fluid velocity passing the drop, cm/s
W - Mass concentration, g/cm
3
W - Cumulative mass concentration, g/cm
W - Total mass concentration, g/cm3
3
lnog
dp - ln dpn
2 in og
Greek
e - Porosity, dimensionless
e - 8.86 x 10" ^ coulomb cm/cm2 - Volt
o
e - Dielectric constant of the particle
n = Single fiber efficiency or single drop
collection efficiency, fraction
£ - Fiber length, cm
yG - Gas viscosity, poise
Pr - Gas density, gm/cm3
pT - Liquid density, g/cm3
LJ
p - Particle density, gm/cm
3
XI1
-------
ABBREVIATIONS AND SYMBOLS (continued)
o - Standard deviation, dimensionless
&
6 - Characteristic diameter, ym
AP - Pressure drop cm W.C.
Abbreviations
ACFM - Actual cubic feet per minute
CDF - Cumulative distribution function
CFM - Cubic feet per minute
GPM - Gallons per minute
SDB - Screen diffusion battery
W.C. - Water column
Xlll
-------
ACKNOWLEDGEMENT
A.P.T., Inc. wishes to express its appreciation for ex-
cellent technical coordination and for very helpful assistance
in support of our technical effort to Mr. Dale L. Harmon, EPA
Project Officer.
xiv
-------
SECTION 1
INTRODUCTION
Air Pollution Technology, Inc. (A.P.T.), in accordance with
EPA Contract 68-02-1496, Experimental Tests of Novel Fine Particu-
late Control Devices, conducted performance evaluations of four
novel fine particulate control devices. They were the Johns-
Manville Cleanable High Efficiency Air Filtration system (CHEAP),
the Air Pollution Systems Electrostatic Scrubber and Electro-
Tube, and the TRW Charged Droplet Scrubber.
The performance evaluations included inlet and outlet particle
sampling measurements with cascade impactors and the A.P.T. Screen
Diffusion Battery. Mathematical performance models were developed
for the CHEAP and the Electrostatic Scrubber. The results of the
performance evaluations are presented in the text.
The effects of charged particles on cascade impactor data
were established as part of this contract because several of the
performance evaluations involved sampling charged particles.
The charged particle experiments will be covered in detail in a
separate report.
-------
SECTION 2
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
Field measurements on four control devices were carried
out to determine the collection efficiency as a function of
particle diameter. A summary of the performance tests is
given in the tabulation below.
Control Device
CHEAP
( Johns -Manville)
Electrostatic Scrubber
(A.P.S.)
Electrotube
(A. P. S.)
Charged Droplet
CT.R.WO
Source
Diatomaceous earth
calciner and dryer
Ti02 test aerosol
Ti02 test aerosol
Coke oven
Pres-
sure
Drop
cm W.C.
48-53
40
3
10
Perfor-
mance
Cut Dia.
ymA
0.8
0.35
<0.1
0.35
Mathematical models were developed for the CHEAP and for
the electrostatic scrubber. The models give reasonable per-
formance predictions. The electro-tube is similar to a wetted
wall electrostatic precipitator, for which a model is available
in the literature. TRW has developed a model for their own
scrubber.
CONCLUSIONS
The program achieved the principal objective of obtaining
reliable performance data for four novel devices. Based on the
data obtained, the following conclusions can be drawn.
-------
1. The CHEAP appears to be a reasonably efficient device
for fine participate control of saturated emissions from a
diatomaceous earth calcining and drying process. Particle
growth occurring in the pre-cleaner prior to the CHEAP is
beneficial to the fine particle collection efficiency of the
system.
The CHEAP, with an approximate actual (wet) cut diameter
of 0.8 ymA and a pressure drop of 50 centimeters W.C., is
comparable to a venturi with a pressure drop of 20 centimeters
W.C.
2. Experimental test results for a Ti02 test aerosol,
which has a dispersed mass median diameter of approximately 1.0
ymA and a standard deviation of 2.2, for the Air Pollution
Systems electrostatic scrubber show overall collection effi-
ciencies ranging from 86% to 974 (average 92%) with the ionizer
on for a pressure drop across the system of 43 cm W.C. With
the ionizer off, the electrostatic scrubber becomes a conven-
tional venturi with efficiencies from 661 to 90% (avg 82%),. 45cmW.C.
Using the experimental data and the average overall pene-
trations, effective cut diameters (i.e., the cut diameter for
a conventional high energy scrubber with the same overall
penetration) for the scrubber with the ionizer on and off
were determined. The effective cut diameter computed for the
ionizer on ranged from 0.23 to 0.54 ymA and from 0.42 to 0.66
ymA for the ionizer off.
3. Experimental tests of the A.P.S. Electro-Tube were
done with Ti02 aerosol which had a mass median aerodynamic
diameter of 1.2 ymA and a standard deviation of 2.2. Overall
collection efficiencies ranged from 96.9% for high gas flow
rates (22.9 Am3/min) to 99.3% for low gas flow rates (16.9
AmVmin).
Experimentally determined penetrations of 0.5 ymA aero-
dynamic diameter particles were 8.6%, 3.71, and 1.1% for high,
medium, and low gas flow rates, respectively. For 1.0 ymA
aerodynamic diameter particles the penetrations were 4.3%,
2.3%,. and 0.68% for high, medium, and low gas flow rates!
respectively. For 2.0 ymA aerodynamic diameter particles'
3
-------
the penetrations were 0.22%, 0.39%, and 0.38% for high, medium,
and low gas flow rates, respectively.
The Electro-Tube performance was found to be similar to
that which can be achieved in small wet electrostatic pre-
cipitators with the same ratio of plate area to volumetric
flow rate.
4. Experimental test results on the TRW Charged Droplet
Scrubber showed overall collection efficiencies of 94.1% for
low gas velocity and low electrode voltage, 88.2% for high
gas velocity and high electrode voltage, 88.3% for high gas
velocity and low electrode voltage, and 94.9% for low gas
velocity and high electrode voltage.
The TRW Charged Droplet Scrubber was compared to a
conventional venturi type scrubber operating at the same
overall efficiency for the same inlet particle dust mean
diameter (d ) and standard deviation (a ). The power consump-
Po o
tion for the TRW Charged Droplet Scrubber was between 17.0
and 25.9 W/(m3/min) for the electrical section and 33.3 W/(m3/min)
for the fan power based on a fan and moj:or efficiency of 50%
and a system pressure drop of 10.1 cm W.C. The power consump-
tion for a conventional venturi type scrubber at the same
overall efficiency as the TRW system was between 354 and 714
W(m3/ndn) for a 50% fan and motor efficiency. Therefore,
based on power consumption the TRW Charged Droplet Scrubber
has an advantage over a conventional venturi type scrubber.
Based on 8,000 hours operation per year and a cost of
3
-------
submicron particles.
Three of the devices tested, the APS Electrostatic Scrubber
and Electro-Tube and the TRW Charged Droplet Scrubber, had
charged particles at the exit. Therefore, the performance data
obtained are expected to differ somewhat from those based on
theoretical cascade impactor performance.
RECOMMENDATIONS
The primary goal of this work was to make performance
evaluations of several novel devices for controlling fine par-
ticulate emissions.
Several of the novel devices tested showed a savings in
power consumption for a given level of control. The capital
costs for the equipment is not well documented at this time
since the devices tested were either in the laboratory or
pilot plant stage. The choice of a control device for removing
fine particulates depends on the following:
1. Size distribution and chemical composition of the
particulate.
2. Gas flow rate, temperature, and composition.
3. Annualized cost of the control equipment which meets
the performance requirement.
The performance characteristics of the control devices
reported here are only for one type of particulate in a labora-
tory or pilot plant system. The fourth scrubber was used on a
process where solution-induced condensation could augment its
efficiency. Therefore the information contained in this report
can only be used as a guide in considering any of these devices.
It is recommended that more information pertinent to the case
at hand be obtained and perhaps a pilot scale test be run before
final selection of any of these novel devices for removing sub-
micron particulates from stack gases.
-------
SECTION 3
CHEAP
SOURCE AND SCRUBBER
The CHEAP (Cleanable High Efficiency Air Filter) is
primarily a wetted fibrous bed system. It consists of water
sprays to wet and clean the filter medium, a rotary drum
containing a fibrous "sponge" filter medium and a water bath
reservoir for cleaning the rotary filter. Particle collec-
tion is accomplished by filtration mechanisms in the rotary
filter.
The CHEAP unit tested is installed on a diatomaceous
earth calciner and dryer. Figure 1 is a schematic drawing
of this system. Prior to the CHEAP, there is a cyclonic
pre-cleaner with water sprays for removing large particles
in the inlet stream. Emissions enter the pre-cleaner,
saturated at a temperature of approximately 75°C, where they
are acted upon by water sprays and centrifugal forces which
collect the large particles in the stream. The gas exiting
the pre-cleaner, saturated at 63°C, then enters the CHEAP.
Water sprays again contact the gas as it is drawn through
the rotating filter drum where the final cleansing action
takes place.
The gas then leaves through the end of the drum while
the particles are washed from the filter media as the drum
rotates into the water bath reservoir. With the help of
two blowers, the gas is forced up the stack and flows into
the atmosphere as a saturated plume at approximately 60°C.
The particle laden water in the reservoir is periodically
drained into the plant's main water purification system
and then refilled.
6
-------
OUTLET SAMPLE
PORTS
STACK
INLET SAMPLE PORTS
DUAL BLOWER
UNIT
WATER
SPRAYS
ROTATING
FILTER
DRUM
WATER
LEVEL
DRAIN
S\:
(
/«x-
STRAIGHTENING
VANES
SAMPLE PORT FOR
PARTICLE GROWTH
pTE£TS
TO ROTOCLONE
CYCLONIC
PRECLEANER
DIATOMACEOUS
EARTH CALCINING
AND DRYING PROCESS
CHEAP UNIT
Figure 1. Schematic of CHEAP unit.
-------
TEST METHOD
The performance characteristic of the CHEAP was determined
by measuring the particle size distribution and mass loading
of the inlet and outlet gas sample simultaneously.
The performance tests consisted of two test series and
are reported in greater detail by Calvert et al. (1975 ).
The first series were performed in November 1974. Investiga-
tion of particle growth in the cyclonic pre-cleaner was the
objective of the second test series during March 1975.
OPERATING CONDITIONS
The CHEAP operating conditions during the first test
period were as follow:
1. Gas flow rates and related parameters as shown in
tabulation below:
DUCT
Temperature
Velocity
Am3/min
ACFM
DN mVmin @ 0°C
DCFM @ 70°F
Vol. IHaO vapor
Pressure
INLET
63°C
12.2 m/s
710
25,000
480
18,300
17
-7.6 cm W.C.
OUTLET
60°C
5.8
Same
Same
Same
Same
0.2
m/s
as
as
as
as
17
cm
inlet
inlet
inlet
inlet
W.C.
2. Water flow rate to the CHEAP system was reported
as approximately 0.053 m3/min (14 GPM).
3. Pressure drop through the CHEAP system was
approximately 48-53 cm W.C. (19-21 in. W.C.) during the
test period.
4. The L/G ratio during the test period was approximately
0.11 £/m3.
DATA
Particle Data
A total of 14 simultaneous inlet and outlet sampling runs
were conducted during the first series of tests on the CHEAP
-------
system. One-point sampling was employed for all runs except
Runs 13 and 14, which were EPA Method 5 tests.
The data were grouped into four data sets. The first
two data sets (runs 4 and 5 and runs 6, 7, and 8) were
obtained at identical one-point sampling locations and
identical plant conditions. The third set of data points
(runs 9 through 12) was obtained at the previous sample
locations but the plant was switched from an oil-fired to a
gas-fired drying process. Finally, the fourth data set
was obtained from two EPA Method 5 tests (runs 13 and 14).
These data sets are presented in detail by Calvert et al.
(1975 ).
As seen in Figures 2 and 3 the following aerodynamic mass
median diameters and standard deviations were found:
RUN NO.
6, 7, 8
9, 10, 11, 12
INLET
dpg,ymA ag
0.82 4.2
0.82 3.9
OUTLET
dpg,ymA
0.82
a
g
3.9
Particle Growth
Size distributions for particle growth, runs 18 and 19,
are given in Figures 4 and 5 , in which the particle size
(wet or dry) is plotted against the cumulative mass percentage
of dry solids. The amount of particle growth is then related
to the difference between the two curves at each mass percent
solids. These figures show that the small particles grow
proportionately more than the large ones, as would be the
case if the particles acted as condensation nuclei with a
small fraction of soluble material in a super-saturated gas.
The aerodynamic mass median diameter of the dry particle
size is almost doubled after particle growth occurs. Aero-
dynamic mass median diameters and standard deviations for
runs 18 and 19 are given below:
-------
10.0
5.0
O RUN #6 I
A RUN #7
| RUN #8
.
2 5 10 20 40 60 80 90
MASS PERCENT UNDERSIZE
Figure 2. Inlet size distributions.
10.0
5.0
1.0
-
0.5
0.4
0.3
0.2
0.1
2 5 10 20 40 60 80 90 98
MASS PERCENT UNDERSIZE
Figure 3. Inlet and outlet size distributions.
-------
A Run 18 Wet Size
Q Run 18 Dry Size
5 10 20 50 80 90
MASS PERCENT UNDERSIZED
Figure 4. Run 18 wet and dry size distributions.
5 10 20 SO 80 90
MASS PERCENT UNDERSIZE, S
Figure 5. Run 19 wet and dry size distributions.
-------
DRY PARTICLE SIZE
RUN NO. d , umA a
Pg g
18 1.4 3.2
19 1.1 3.1
WET PARTICLE SIZE
d , umA o"
Pg g
2.2 2.3
2.1 2.0
Opacity
Plume opacity of the CHEAP system averaged 10% during the
test period. The visual observation method was used for all
opacity measurements which were taken by a Johns-Manville
employee who was a certified observer trained in a California
Air Resources Board "Smoke School". According to the observer,
visible measurements were taken on a hill above the stack
approximately fifteen meters away. A detached plume enabled
the observer to read opacity at the stack.
PARTICLE PENETRATION
Particle penetration versus particle aerodynamic diameter
was calculated from simultaneous cascade impactor runs. Since
the cascade impactors were heated, the particle diameter mea-
sured was that of a dry particle.
Particle penetration versus dry particle for runs 10, 11,
and 12 is plotted in Figures 6, 7, and 8. The penetration
curves are flat and show a relatively constant penetration of
SI for all particle sizes.
In the scrubber, the particles are wet which gives the
submicron particles a larger aerodynamic diameter than their
physical size would. The wet particle size was not measured
during the first test period.
During the second test period, the CHEAP unit was shut down.
Therefore, it was impossible to measure the wet particle diameter
in the scrubber inlet and outlet simultaneously. However, the
wet and dry particle size distributions were determined at the
CHEAP inlet by means of simultaneous heated and unheated cas-
cade impactor sampling runs. The solution induced particle
12
-------
.10
l/J
.-:
:
s
c
-
...
,-
,-
c
-
.-
-
•-
o.
.05
.03
.01
.10
0.3 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER,
3.0
-
u
S
.05 g
:
S
-
--
I
B
—
.
-
.03 s
0.3 0.5 l.o 2.0 3.0
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 6. Penetration versus dry particle diameter
for Run 10-
Figure 7. Penetration versus dry particle diameter
for Run 11 .
-------
.10
o
i-i
H
U
§
.05
.03
w
ex,
w
CJ
1—t
H
.01
^
0.3 0.5 1.0 2.0 3.0
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 8 . Penetration versus dry particle diameter
for Run 12 .
14
-------
growth can be determined from the wet and dry particle size
distribution data.
The inlet dry particle size distribtuions are comparable
for the first and second test periods. Therefore, by assuming
the outlet is also comparable and the extent of particle growth
in the outlet is the same as inlet, we can calculate the
grade penetration curve for wet particles from the dry particle
penetration curve and from the relation between wet and dry
particle size distribution curves. The result is shown in Figure 9.
Particle growth appears to have a beneficial effect
on particle penetration, especially in the submicron range.
Particles in this range grow to as much as 3 times their orig-
inal diameter. They, in turn, are more easily captured by
inertial impaction and thus a lower particle penetration occurs.
Instead of particle penetration increasing for submicron par-
ticles, it remains constant due to the particle growth and
thus the penetration curves appear flat.
The performance cut diameter of the CHEAP calculated by
equivalent cut diameter method is 0.8 ymA wet.
MATHEMATICAL MODELING
The prediction of particle penetration for the CHEAP
required the development of mathematical models because none
were available from the literature. We can mathematically
represent the filter as a fiber bed consisting of an array
of equally spaced cylinders. The "Scrubber Handbook" (Calvert et al
1972) gives the following equation for the prediction of
particle penetration of a clean fibrous bed:
Pt(dp) = exp [- S n] - exp [- 4 ^\^ n] U)
where "n" is the effective collection efficiency of a single
fiber in the bed for all collection mechanisms.
By assuming negligible interaction among fibers, the col-
lection efficiency of a fiber in a bed can be approximated by
the collection efficiency of an isolated fiber. This assump-
tion might slightly underestimate the fiber efficiency but
15
-------
n.i
_\ I I I I l
o
•H
+J
O
o
H
W
z
UJ
0.05
0.01
0. 3
1 ~
WET PARTICLE —
DRY PARTICLE
I lilt
I
1 .0
PARTICLE DIAMETER, umA
4.0
Figure 9. Experimental grade penetration
curve for CHEAP.
16
-------
the fibers are not all oriented normally to the gas flow
direction and this will lower the collection efficiency.
The pressure drop across the fiber bed is the sum of the
drag losses of all fibers. We used the drag coefficient for
an isolated cylinder, as is consistent with the assumption of
negligible interaction among fibers.
£(l-e) pr Cn u3.
AP = 6.5 x 10"" —t!—k
df
Equations (1) and (2) were used to predict the performance
of the CHEAP. The result is shown in Figure in, a plot of cut
diameter versus pressure drop for various fiber diameters. The
cut diameter versus pressure drop relation is highly dependent
on the diameter of the fiber but not much on the solidity fac-
tor. The circle in the figure represents the data we deter-
mined experimentally. The fiber diameter and the porosity of
the filter medium of the scrubber were not disclosed as they are
proprietary. Therefore, it was not possible to determine how
well the model predicts the performance.
It is possible to compare our model with the data of Rei
and Cooper C1976) for tests on a pilot scale unit of the CHEAP.
They reported the volume fraction void of the filter medium to
be 97% and the fiber diameters to be 64 ym, 44 ym, and 36 ym
for foams with 18, 26, and 33 pores per cm. They also reported
the measured cut diameter was somewhere below 0.5 ymA for pres-
sure drops, ranging from 40 to 90 cm W.C. The dashed line in
Figure 10 shows their data, which are consistent with our
predictions.
ECONOMICS AND OPERATING PROBLEMS
Cost data for the CHEAP were not provided by the
manufacturer. According to the manufacturer, the technology
is not yet sufficiently well established to provide reliable
capital and operating costs.
The CHEAP, although a temporary installation, operated
very smoothly during the testing period; however, plant
17
-------
3.0
1.0
UJ
2:
<
HH
a
CJ
0.1
= 300 ym
0 A.P.T. DATA
RE I AND COOPER'S DATA
I
I
L I
I I till
10 100
PRESSURE DROP, cm W.C.
300
Figure 10. Cut diameter versus pressure drop for
fibrous bed.
18
-------
process shutdowns delayed testing on schedule. The major
problem in this CHEAP installation is operating under a
corrosive atmosphere. Residual chlorides from the diatomaceous
earth process and sulfates from the oil-fired furnace, together
with a saturated gas stream, tend to accelerate corrosion on
the internal parts of the CHEAP.
Carbon steel fans in both of the blowers were gradually
eroding, causing an imbalance in the units. This, in turn,
caused considerable vibrational problems and excessive wear on
the motors' main shaft bearings.
CONCLUSIONS
The CHEAP appears to be a reasonably efficient device for
fine particulate control of saturated emissions from a diato-
maceous earth calcining and drying process. Particle growth
occurring in the pre-cleaner prior to the CHEAP is beneficial
to the fine particle collection efficiency of the system to a
certain extent.
According to the test results obtained on the CHEAP (Clean-
able High Efficiency Air Filter) system, particle penetration
is relatively independent of dry particle size. Penetration
is approximately 5% with the mean dry particle diameter equal
to 0.82 ymA (1-5 ymA wet). The performance of the CHEAP, with
an approximate actual (wet) cut diameter of 0.8 ymA and a
pressure drop of 50 centimeters W.C., is comparable to a ven-
turi with a pressure drop of 20 centimeters W.C. However,
according to the manufacturer the initial capital investment
for the CHEAP is much less.
19
-------
SECTION 4
A.P.S. ELECTROSTATIC SCRUBBER
SOURCE AND CONTROL SYSTEM
The pilot scale "Electrostatic Scrubber" of Air Pollu-
tion Systems is basically an electrostatic charger (or
ionizer) followed by a venturi scrubber. Figure 11 is a
schematic diagram of the pilot system. An electrode is
placed upstream of the venturi to charge the inlet particles,
which then enter the venturi throat. The gas stream atomizes
the central water spray in the venturi throat and the charged
particles, according to A.P.S., are then attracted and col-
lected by the highly polarized water molecules.
The charged particles are also collected on the walls
of the ionizer section prior to the throat of the venturi.
A thin film of water is run down the inclined surfaces to
keep the walls clear and prevent high voltage arcing. The
particle laden water droplets are then collected by a
cyclonic separator and sent into a settling tank (clarifier).
The water can then be recycled back into the scrubber system.
However, during the test program, fresh water was used.
The ionizer consists of an electrode supported in the
inlet of the venturi section. According to Air Pollution
Systems, a stable electrical discharge of high intensity
is maintained across the venturi throat between the center
electrode and the wall. A.P.S. claims that the average field
that can be maintained across the electrode gap (space between
the electrode probe and the wall) is substantially higher,
14-16 kV/cm, than that of a standard electrostatic precipitator,
4 kV/cm,
The pilot scrubber had a maximum capacity of 28 Am3/min
(1,000 ACFM) at 16°C. Under testing conditions the scrubber
20
-------
TO INDUCED
^^^ DRAFT PAN
CLEAN
GAS
PITT
VENTURI
SPRAY
WATER
TO WASH
IONIZER
WALL
IONIZER
SECTION
HIGH
VOLTAGE
POWER
SUPPLY
CYCLONE
ENTRAINMENT
SEPARATOR
CLARIFIER
Figure 11. APS electrostatic scrubber.
21
-------
was operated at 21.0 Am3/min (740 ACFM) @16°C and at 22.7
Am3/min (800 ACFM) @ 16°C. Liquid flow rates (total) were
set at 30.3 i/min and 41.7 H/min during the test program.
L/G ratios were 1.4 £/m3 and 1.8 £/m3, respectively, for
the above flow rates. Electrode probe cooling air was
introduced at a rate of 1.1 m3/min (40 CFM). The addition
of the cooling air was accounted for as dilution air in the
outlet sample calculations.
A 3-horsepower induced draft fan with a capacity of 28
Am3/min at a pressure of 45 cm W.C. was located downstream of
the outlet sample ports. Gas flow rates were adjusted
through a damper upstream of the fan.
The test aerosol used during this study was titanium
dioxide, TiOa, which has a density of about 3.0 g/cm3, as
dispersed. The mass median diameter of Ti02 aerosol was
approximately 1.0 microns aerodynamic, ymA, as dispersed,
with a standard deviation of 2.2.
TEST METHOD
The performance tests were completed during two test
periods. For the first tests performed in June 1975,
the University of Washington Mark III (U.W.) cascade impac-
tors were used for particle measurements above 0.3 ymA in
diameter. Measurements were conducted simultaneously in
the inlet and outlet ducts.
An Air Pollution Technology portable screen diffusion
battery (A.P.T.- SDB ) was used for particle measurements
from 0.1 ym to 0.01 ym (actual). During an impactor run,
several inlet and outlet fine particle size measurements
were taken alternately with the portable diffusion battery.
Since the system remained fairly constant during each run,
alternate inlet and outlet SDB measurements were con-
sidered to approximate simultaneous sampling.
Additional tests were performed in July 1975 to determine
whether static charges carried by the particles was affecting
the particle classification of the cascade impactor. Polonium
22
-------
210 charge neutralizes (Staticmaster, Ionizing Unit Model
No. 2U500) were connected prior to the impactors on both
the inlet and the outlet.
In the first tests, particles form the particle gen-
erator were not neutralized. In order to check the possi-
bility that particle charge might affect scrubber perfor-
mance when the ionizer is off, a charge neutralizer was
added to the dust disperser. A two foot glass tube con-
taining six Staticmasters was placed on the end of the
particle generator during the second tests.
OPERATING CONDITIONS
A total of thirty-one simultaneous inlet and outlet sample
runs were performed on the electrostatic scrubber. The data
are grouped into four data sets.
During the first three series of tests (data sets A, B,
and C) the scrubber was tested with the ionizer on and off
and during the fourth series (data set D) with the ionizer
on.
The scrubber was tested at 21 An3/min C740 ACFM) at 16°C
for the first series of tests (runs 1-10, data set A). The
pressure drop across the venturi was 40 cm W.C. with liquid
flow rate? to the ionizer wall wash and the venturi throat
set at 7.6 l/min and 22.7 SL/min, respectively.
The second series of tests (runs 11-26 , data set B)
were run at 22.7 AmVmin (800 ACFM) at 16°C with the same
pressure drop across the venturi of 40 cm W.C. The liquid
flow rate was increased to 34.1 5,/min at the venturi throat
and the ionizer wall wash remained the same at 7.6 £/min.
An overhanging gasket and water feed tube were removed from
the scrubber throat between sets "A" and "B".
The third series of tests (runs 27-30, data set C) were
run at identical operating conditions as the second series.
These runs were taken to determine the effect of the static
charge induced by the aerosol generator. A cross-shaped
baffle was introduced into the test aerosol mixing duct prior
to set "C", in order to cause better mixing of air and the
23
-------
aerosol stream from the dust blower.
Finally the last series of runs attempted, used alumina as
a test aerosol instead of titanium dioxide. Because of particle
generator plugging, run 31 (data set D) was the only test performed
on this aerosol.
Gas flow rates, water flow rates and other related operating
parameters are tabulated for each data set and are given in the
Tables 1 and 2.
DATA
Cascade Impactor and Diffusion Battery Data
The inlet and outlet size distributions obtained during the
performance tests are given in Table 3 . These were determined
from cascade impactor and diffusion battery runs which are
explained in greater detail hy Calvert et al. (1976a). A
summary of the mass loadings and overall penetrations is
given in Table 4 .
PARTICLE PENETRATION
Particle penetration versus particle size was plotted for
the data obtained from the electrostatic scrubber. A composite
of the runs in data set "A" is represented in Figure 12, which
shows the effect of the ionizer on penetration. Penetrations
for the individual runs in sets "B" and "C" are shown in
Figures 13 and 14. Figure 15 shows composite curves for sets "A"
"B", and "C" for overall comparison.
As can be seen in Figures 12 through 15 the ionizer does
improve the scrubber performance in all cases. However, the
amount of improvement is least in data set "C", where the
most precautions were taken to neutralize and mix the test
aerosol entering the scrubber.
The penetrations determined from diffusion battery data
are presented in Figures 16 through 19 along with the individual
penetrations for the cascade impactor runs of data set "C".
Particle density variation and loading fluctuation account
for discrepancies between impactor and D.B. penetrations.
24
-------
TABLE 1. DATA SET A, RUNS 1-10
CONDITION INLET OUTLET
Temperature
Velocity
AmVmin @16°C
ACFM @16°C
DNmVmin @ o°C
DSCFM @21°C
Vol. % H20 vapor
Static Pressure
15-21°C
8.4 m/s(27.6 ft/s
21.0
740.0
19.6
745.0
1.1
-0.6 cm W.C.
Pressure drop across venturi
Pressure drop across separator
Ionizer wall wash flow rate
Venturi water flow rate
*Electrode cooling air
L/G ratio
14-20
S.I m/s (16
22.1
780.0
19.7
748.2
1.6
-43.2 cm
40.0 cm W.
2.5 cm W.
7.6 H/min
22.7 Z/min
1.1 m3/min(40
1.4 l/m*
°C
.7 ft/s)
W.C.
c.
c.
CFM)
*Electrode cooling air introduced into system and
accounted for as dilution air.
TABLE 2. DATA SETS B, C AND D, RUNS 11-31
CONDITION
Temperature
Velocity
AmVmin @16°C
ACFM @16°C
DNmVmin l?0°C
DSCFM @21°C
Vol . % H20 vapor
Static Pressure
INLET OUTLET
15-21°C
7.5m/s (24.5 ft/s
22.7
800.0
21.2
805.0
1.1
-0.6 cm W.C.
Pressure drop across venturi
Pressure drop across separator
Ionizer wall wash flow rate
Venturi water flow rate
"Electrode cooling air
L/G ratio
14-20°C
5.4m/s (17.8 ft/s
23.8
840.0
21.2
806.0
1.6
-43.2 cm W.C.
40.0 cm W.C.
2.5 cm W.C.
7.6 i/min
34.1 4/min
1.1 m'/min (40 CFM)
1.8 i/m3
*Electrode cooling air introduced into system and
accounted for as dilution air.
25
-------
TABLE 3. INLET AND OUTLET SIZE DISTRIBUTION SUMMARY
Run Set
No.
Z A
3
4
5
6
11 B
13
14
16
18
20
2]
22
23
24
25
26
27 C
28
29
30
31 D
Ionizer
ON
ON
ON
OFF
OFF
ON
ON
OFF
OFF
ON
ON
ON
ON
ON
ON
OFF
OFF
OFF
OFF
ON
ON
ON
INLET
dpg, umA
0.95
0.95
0.95
1.1
1.1
0.90
0.90
1.0
1.2
1.2
1.1
0.94
0.94
0.94
0.97
1.2
1.5
1.4
1.8
2.1
1.9
°g
2.2
2.2
2.2
2.3
2.3
2.0
2.0
2.3
2.3
2.3
1.9
1.9
1.9
1.9
2.2
2.1
2.0
1.8
1.9
2.1
Z.3
OUTLET
dpg. u"iA
0.90
0.90
0.90
1.0
1.0
0.83
0.83
0.95
0.95
0.90
0.90
1.1
0.92
1.3
1.1
1 .1
1.4
1.2
CTK
1.7
1.7
1.7
1.9
1.9
1 .6
1 .6
1.8
1.8
1.7
1.7
1.8
1.8
1.8
1.9
1.5
1 .5
2.1
to
TABLE 4. TEST RUN SUMMARY
DATE
6/17
6/18
6/18
6/18
6/19
6/19
6/19
6/20
6/20
6/20
6/23
6/23
6/23
6/24
6/24
6/24
6/25
6/25
6/25
6/25
6/26
6/26
6/26
6/27
6/27
6/27
7/17
7/17
7/17
7/17
7/18
RUN
NO.
Ifspt A"}
^ O C I rt J
2 "
3 "
4 "
5 "
6 "
7 "
8 "
9 "
10 "
LI (set B)
12 "
13 "
14 "
15 "
16 "
17 "
18 "
19 "
20 "
21 "
22 "
23 "
24 "
25 "
26 "
27(set C)
28 "
29 "
30 "
31(set D]
SAMPLE
DEVICE(l)
Inlet Outlet
F (21 B
1 V. *• 1 "
I I
I I
I I
I I
I I
F F
F F
F F
F B
I I
F B
I I
I I
F F
I I
F F
I I
F,D B
I I
I ,D F
I,D • F
I,D F
I I,D
F I,D
I,D F
I,D (3) I,D
I,D I,D
I,D I,D
I,D I.D
1(4) I
IONIZER
ON
ON
ON
ON
OFF
OFF
OFF
OFF
ON
OFF
ON
ON
ON
OFF
OFF
OFF
ON
ON
ON
ON
ON
ON
ON
ON
OFF
OFF
OFF
OFF
ON
ON
ON
MASS LOADING
mg/DNm3
Inlet
146. S
156.1
188.5
524.6
396.1
185 .1
927.9
1081.2
218.7
152.7
132.0
153.4
216.9
594.2
516.9
472.4
495.9
69.4
192.2
48.1
53.7
100.1
85.1
153.9
102.6
301 .8
161.1
257.0
91.0
183.7
Outlet
1 n
1 . U
12.8
12.0
16.1
75 .2
62.1
62 .4
108.2
37.7
41.8
10.3
13.8
11 .4
40.8
79.2
65.1
24.2
18 .4
8 .4
12.5
8.1
7 .6
10.3
12.7
31 .8
21.3
30 .6
24 .0
7.9
5.0
10 .9
Notes:
(1)
(2)
(3)
(4)
I = Impactor run F =
B = Blank impactor run
Total filter run
D = Diffusion battery run
Faulty filter holder, no inlet data obtained
Runs 27-31, charge neutralizers connected prior to
impactors and also connected on the dust feeder.
A cross was installed downstream of the dust feeder
to promote mixing.
Alumina used as test aerosol
-------
ts)
1.0
0.5
f-i
o
2
u.
1.0
g
ri]
g 0.05
o.
0.01
I
IONIZER
ON
i i I i i I
10
0.2 0.5 1.0 5
PARTICLE DIAMETER, pmA
Figure 12. Penetration vs. particle diameter
for data set A.
1.0
0.5
z
o
E-
(J
g 0.1
f-
e:
z 0.05
tu
o.
0.01
0.2
X \ IONIZER OFF
I I i i i
0.51.0 5
PARTICLE DIAMETER, UmA
10
Figure 13. Penetration vs. particle diameter
for data set B.
-------
CO
1.0
0.5
2
O
3
u.
1.0
O
HH
f-c
<
a:
H
fM
z 0.05
u
PU
0. 01
IONIZER OFF
0.2 0.5 1.0
PARTICLE DIAMETER, ymA
10
Figure 14. Penetration vs. particle diameter
for data set C.
1.0
0.5
O
H
a,
u.
0. 1
0.05
0.01
IONIZER OFF
0.2 0.5 1.0 5
PARTICLE DIAMETER, pmA
10
Figure 15. Penetration vs. particle diameter
for data sets A, B, 5 C.
-------
o
an
u,
tu
UJ
D.
1.0
0.5
0.1
0.05
0.01
'T
DB
CI
I I II 1 jj
0.05 0.1 0.5 1.0 5.0
PARTICLE DIAMETER, umA
Figure 16. Penetration vs. particle
diameter using both diffu-
sion battery and cascade
impactor, run 27 (set C,
ionizer off)
c
t—i
L;
Li,
z
o
I—'
E-
E-;
Z
CL.
0.5
0. 1
0.05
0.01
I I IT
DB
CI
0.05 0.1 0.5 1.0
PARTICLE DIAMETER, umA
5.0
Figure 17. Penetration vs. particle
diameter using both diffu-
sion battery and cascade
impactor, run 28 (set C,
ionizer off)
-------
o
I—I
E-
U
<
oc
U-
tu
1.0
0.5
0.1
u. g 0.05
O O.
0.01
11
i n_
DB
0.05 0.1 0.5 1.0 5.0
PARTICLE DIAMETER, ymA
Figure 18. Penetration vs. particle
diameter using both diffu-
sion battery and cascade
impactor, run 29. (set C,
ionizer on)
u
<
ac
u.
c
H
1.0
0.5
0.1
pu
g 0.05
D.
0.01
DB
0.05 0.1 0.5 1.0 5.0
PARTICLE DIAMETER, umA
Figure 19. Penetration vs. particle
diameter using both diffu-
sion battery and cascade
impactor, run 30. (set C,
ionizer on)
-------
The penetration curves indicate an increase in efficiency
for particles smaller than the 0.1 to 0.2 ymA range. From
these results and published literature, it is evident that
smaller, more highly diffusive particles experience increased
collection efficiency.
MATHEMATICAL MODEL
The prediction of particle penetration for the A.P.S.
Electrostatic Scrubber required the development of a math-
ematical model because none was available from either A.P.S.
or the literature. The following section presents the deriva-
tion of a predictive model by means of the unit mechanism
approach. While the form of the equation is the consequence
of theoretical relationships, an empirical constant, f, is
included and its value was assumed to be the same as for non-
charged venturi scrubbers.
In a venturi scrubber the most important unit mechanism
responsible for particle collection is the collection by drops
and the predominant collection phenomenon is inertial impac-
tion. When particles are charged, then in addition to the
inertial force, electrostatic forces are present that force
the particle towards the drop, i.e., increase the collection
efficiency of the drop.
Calvert et al. (1973) calculated the theoretical increase
in collection efficiency by flux forces. They relate the
single drop collection efficiency to the inertial parameter
as shown in Figure 20 with flux deposition number, N-j,, as
parameter. Inertial impaction parameter and flux deposition
number, NFD, are defined by the following equations:
C'Vodp2
S = ~^
where:
_ UF _ Particle electrostatic deposition velocity
FD ~ u " ~ Fluid velocity past sphere
31
-------
1x10
- 2
10-
INERTIAL PARAMETER, K
Ficjure 20. Efficiency of single drop versus inertial parameter
Npori - 9-6 with N™ as parameter
Red
-------
K = inertial impaction parameter, dimensionless
C' = Cunningham slip factor, dimensionless
p = particle density, g/cm3
d = particle diameter, cm
yG = gas viscosity, poise
d, = drop diameter, cm
UQ = fluid or gas velocity passing the drop, cm/s
From Stokes' law, the particle drift velocity is given by
C' Q E
where E = field strength, kV/cm
Q = electrical charge carried by the particle,
coulomb
It is assumed that aerosol particles are charged to
saturation after the particles pass through the corona discharge
section. If diffusional charge is neglected (this will give
conservative results) , then according to Oglesby and Nichols
(1970), the saturation charge can be calculated from the
following equation:
Q = 3 IT eo d* Eo (5)
where e = dielectric constant of the particle
e0 = 8.86 x 10 l '* coulomb cm/cm2--volt
EO = charging electric field strength, kV/cm
The atomized drop diameter in a venturi is assumed to be
the Sauter mean diameter, which is predicted by means of the
Nukiyama and Tanasawa correlation
n (6)
-------
where Qj = volumetric flow rate of liquid, m3/min
Q£ = volumetric flow rate of gas, m3/min
ur = gas velocity, cm/s
u
Once fNnr.' and 'K ' are calculated, single drop collection
r U p
efficiency can be read from Figure 21 or from Figure 22. Single
drop collection efficiency 'n' is then related to penetration
based on an equation given by Calvert (1968) for a venturi,
d
k PL dd uct f° ,,
55 Qr yr / n df
IT (i ^
Pt, - » rt. : "" / n « (7)
fa
where Pt , = particle penetration for particle size "d "
d p
uGt = venturi throat gas velocity, cm/s
f = an empirical constant
f = initial value of f ~ 0.5
3.
PT = liquid density, g/cm3
n = single drop collection, fraction
This model was applied to predict the collection efficiency
of the A.P.S. scrubber both with the ionizer on and with the
ionizer off. Since the calculation of particle drift velocity
(equation (4) involves the determination of field strength
distribution, which is unknown, a constant value equal to the
field strength at the gap was used. This assumption will lead
to an optimistic drift velocity.
Figure 22 shows the predicted A.P.S. scrubber performance
along with experimental curves. As can be seen, the model
agrees fairly well with experimental values for large particles
but deviates for small particles. This deviation may be
caused by omission of the diffusion collection mechanism in
the theoretical model.
34
-------
1.2
Figure 21. F.fficiency of a single drop, n, versus Npn with
K as a parameter (NR , = 9.6)
-------
1.0
t-
C_3
Z;
O 0.1
UJ
a,
Pi
0.01
0.005
Predic ted
I I I
I III
Ionizer on
— - Ionizer off
n-?~ i.o 10. n
AERODYNAMIC PART I CLP. DIAMETER, ymA
Figure 22. Exp»erimenta 1 and predicted particle pene-
tration versus particle diameter.
36
-------
SECTION 5
A.P.S. ELECTRO-TUBE
SOURCE AND CONTROL SYSTEM
The pilot scale Electro-Tube of Air Pollution Systems
(A.P.S.) is basically a tube electrostatic precipitator with
a central rod electrode and wetted wall collector. Figure 23
is a schematic diagram of the pilot system. The inlet particles
are charged in a high energy field (12 kV/cm) by a high inten-
sity ionizer at the base of the electrode. The charged par-
ticles then migrate to the wetted wall in the body of the device
in a field of 5-10 kV/cm. A.P.S. indicates that initial satura-
tion charge on the particles is higher than the usual 4-5 kV/cm
for a conventional ESP (electrostatic precipitator) and facili-
tates increased migration in the collecting electric field.
The Electro-Tube tested was a laboratory pilot scale unit.
Under testing condition? the Electro-Tube was operated at gas
flow rates of 16.9, 18.9, and 22.9 Am3/min (596, 668, and 808
ACFM, respectively) at 20°C. Liquid flow rates could range
from 1.7 to 6.4 £/min according to the manufacturer's specifi-
cations, but the liquid flow was held constant at 3.8 t/min
during the entire test period. The tube diameter was 30.5 cm
(12 in.).
The test aerosol for this study was titanium dioxide
(Ti02) which had a single particle density of 4.1 g/cm3. The
mass median aerodynamic diameter of the dispersed aerosol was
about 1.2 pmA, geometric standard deviation 2.2, and agglomerate
density about 3.0 g/cm3.
The particle generator used during the test consists of a
feed auger, intermediate blower, deagglomeration orifice
37
-------
FROM HIGH
VOLTAGE SOURCE
ELECTRODE
AEROSOL
INLET
AEROSOL OUTLET
TO BLOWER
OPTIONAL
SECONDARY
AND TERTIARY
IONIZATION
ZONES
HIGH INTENSITY
IONIZER SECTION
U
—TANK
U
OUTLET
WATER DRAIN
Figure 23. Diagram of A.P.S. Electro-Tube
38
-------
chamber, and main blower. A cross-shaped baffle was placed at
the system inlet to ensure adequate mixing of the aerosol with
the incoming ambient air.
TEST METHOD
All tests were performed using modified EPA type sampling
trains with in-stack University of Washington Mark III (U.W.)
cascade impactors. Greased aluminum substrates were used in
the impactors to prevent particle bounce and minimize wall
losses.
The Air Pollution Technology portable screen diffusion
battery (A.P.T. - SDB ) was used for particle measurements
from 0.01 ym to 0.1 ym (actual).
During an impactor run, several inlet and outlet fine
particle size measurements were taken alternately with the
portable diffusion battery. Since the system remained fairly
constant during each run, alternate inlet and outlet SDB
measurements were considered to approximate simultaneous
sampling.
In-stack filter samples were also taken to obtain total
particulate loadings and overall collection efficiencies of
the system.
Blank impactor runs were performed periodically to ensure
that the greased aluminum substrates did not react with the
stack gases. A blank impactor run consisted of an impactor
preceded by two glass fiber filters run at identical sample
conditions as the actual sampling runs. Total filter loadings
were also obtained during the blank impactor run to furnish
simultaneously inlet and outlet mass concentrations and overall
collection efficiency data.
OPERATING CONDITIONS
The operating conditions for the A.P.S. Electro-Tube are
tabulated in Table 5.
39
-------
TABLE 5. OPERATING CONDITIONS
Condition
Inlet
Outlet
Temperature
Velocity in Electro-Tube
Am9/min 8 22.5°C
ACFM 6 22.5°C
DNmVmin @ 0°C
DSCFM § 21°C (70°F)
Vol % H20 Vapor
Static Pressure at
Sample Ports
18-27°C
4.3-5.8 m/s
(14.1-19.1 fps)
16.9-22.9
596.0-808.2
15.4-20.9
585.0-794.0
1.3-2.4
-0.5 to
-1.5 cm W.C.
18-27°C
4.4-6.0 m/s
(14.5-19.9 fps)
17.4-23.8
615.0-841.0
15.9-21.7
605.0-826.0
1.3-2.4
-2.5 to
-4.6 cm W.C.
Pressure drop across Electro-Tube: 0.53 (low flow) to
0.64 (high flow) cm W.C.
Ionizer wall water wash rate: 3.8 £/min (1 GPM)
40
-------
chamber, and main blower. A cross-shaped baffle was placed at
the system inlet to ensure adequate mixing of the aerosol with
the incoming ambient air.
TEST METHOD
All tests were performed using modified EPA type sampling
trains with in-stack University of Washington Mark III (U.W.)
cascade impactors. Greased aluminum substrates were used in
the impactors to prevent particle bounce and minimize wall
losses.
The Air Pollution Technology portable screen diffusion
battery (A.P.T. - SDB ) was used for particle measurements
from 0.01 ym to 0.1 ym (actual).
During an impactor run, several inlet and outlet fine
particle size measurements were taken alternately with the
portable diffusion battery. Since the system remained fairly
constant during each run, alternate inlet and outlet SDB
measurements were considered to approximate simultaneous
sampling.
In-stack filter samples were also taken to obtain total
particulate loadings and overall collection efficiencies of
the system.
Blank impactor runs were performed periodically to ensure
that the greased aluminum substrates did not react with the
stack gases. A blank impactor run consisted of an impactor
preceded by two glass fiber filters run at identical sample
conditions as the actual sampling runs. Total filter loadings
were also obtained during the blank impactor run to furnish
simultaneously inlet and outlet mass concentrations and overall
collection efficiency data.
OPERATING CONDITIONS
The operating conditions for the A.P.S. Electro-Tube are
tabulated in Table 5.
39
-------
TABLE 5. OPERATING CONDITIONS
Condition
Temperature
Velocity in Electro-Tube
Am3/min @ 22.5°C
ACFM @ 22.5°C
DNrn'/min @ 0°C
DSCFM @ 21°C (70°F)
Vol % H20 Vapor
Static Pressure at
Sample Ports
Inlet
18-27°C
4.3-5.8 m/s
(14.1-19.1 fps)
16.9-22.9
596.0-808.2
15.4-20.9
585.0-794.0
1.3-2.4
-0.5 to
-1.5 cm W.C.
Outlet
18-27°C
4.4-6.0 m/
(14.5-19.9
17.4-23.
615.0-841
15.9-21.
605.0-826
1.3-2.4
-2.5 to
-4.6 cm W.
Pressure drop across Electro-Tube: 0.53 (low flow) to
0.64 (high flow)cm
Ionizer wall water wash rate: 3.8 £/min (1 GPM)
s
fps)
8
.0
7
.0
c.
w.c.
40
-------
DATA
Cascade Impactor Data
Data sets for three gas flow rates were obtained for the
A.P.S. Electro-Tube. As the gas flow was fully developed
one-point sample locations were used for all data points.
Table 6 is a summary of the size distribution of inlet
and outlet mass loading, and overall penetration data for the
various runs made. Table 7 summarizes the total filter loading
tests. The original data obtained during the performance test
is described in greater detail in a report by Calvert et al.
(1976b).
Diffusion Battery Data
Diffusion battery data were taken during the last three
days of testing. The runs were made simultaneously with cas-
cade impactors and numbered accordingly. Runs 14 and 16 were
preliminary tests to determine count and flow adjustment
techniques; therefore, they were not included in the data set.
Performances for runs 17, 18, and 19 were determined
by statistical conversion of inlet-outlet size distribution
to cumulative number concentration and calculating the particle
penetrations in the same manner as the cascade impactor analysis.
Calculations were performed by a computerized FORTRAN program.
Table 8 contains the size distribution summary of the
diffusion battery tests. Note that the sizes given are actual
rather than aerodynamic as with cascade impactor analysis.
This is necessary when reducing data from the diffusion
battery as this instrument classifies particles according to
their physical size. Conversion to aerodynamic diameter for
comparison with impactor analysis requires knowledge of the
density.
The titanium dioxide powder used had a density of 4.1 g/cm3
as measured by a pycnometer. This value agreed closely with
published data for Ti02. The solid unitary particles would have
this value for their density. However, the batch of Ti02 that
was used consisted of many agglomerated as well as single par-
ticles. Most of the single particles and the detectible units
41
-------
TABLE 6. SIZE DISTRIBUTION, MASS LOADING, AND OVERALL
PENETRATION DATA*
Run
No.
1
2
3
4
5
6
7
8
9
10
12
13
14
15
16
18
19
INLET
d . umA
Pg'
Fil
°g
:er
Filter
1.02
1.15
1.47
2.1
2.1
2.4
Filter
1.55
2.4
Filter
Filter
1.30
0.87
1.27
1.11
1.05
1.20
1.25
1.20
2.2
2.0
2.1
2.2
2.0
2.5
2.2
1.9
mg/DNm3
177
242
185
248
571
775
1080
606
662
364
167
389
248
375
738
274
240
OUTLET
dpg,pmA
Fil
a
g
:er
Filter
0.69
0.71
0.70
1.4
1.5
1.5
Filter
0.68
1.5
Filter
Fil
0.82
0.69
0.75
0.74
0.68
0. 79
0.66
1.00
;er
1.7
1.6
1.9
1.4
1.8
1.8
1.5
2.2
mg/DNm3
8.7
11.0
8.5
8.7
13.0
14.5
13.7
19.0
9.1
5.0
4.1
7. 0
4.2
2.5
5.8
7.8
1.2
Overall
Penetration
(*)
4. 9
4.6
4.6
3.5
2.3
1.9
1.3
3.1
1.4
1.4
2. 5
1.8
1.7
0.7
0.8
2.9
0.5
*d
is the geometric mass mean aerodynamic particle
diameter in this tahlp.
TABLE 7. TOTAL FILTER PARTICLE LOADING TESTS
Run
No.
1
2
6
8
9
INLET
Mcum
(mg/DNm3)
177
242
775
606
662
DNm3
0.25
0.28
0.12
0.11
0.067
OUTLET
Mcum
(mg/DNm3)
8.7
11.0
14.5
19.0
9.1
DNm3
0.94
1.01
1.23
1.02
0.83
42
-------
of the agglomerates had diameters in the range of 0.1 to 0.3 ym
as measured from scanning electron micrographs. Therefore,
particles larger than about 0.3 ym were mainly agglomerates and,
because of their irregular shape and possible voids, had a
density less than 4.1 g/cm3. The density of the larger particles
was estimated to be about 3 g/cm3.
TABLE 8. NUMBER BASIS SIZE DISTRIBUTION DATA FOR
DIFFUSION BATTERY TESTS*
Run
No.
17
18
19
INLET
dpN'ym
0.37
0.14
0.34
a
g
7.8
7.2
7.9
OUTLET
dPN'Pm
0,12
0.035
0.19
ag
7.6
4.3
7.5
*d N is the geometric number (count) mean
p particle diameter in this table.
PARTICLE PENETRATION
The overall penetration summary is presented in Table 6
Total mass loadings were taken by cascade impactors or total
filter samples. Overall penetrations tended to decrease with
decreasing volumetric gas flow rates.
Average overall penetrations for the Electro-Tube were
3.1% for 22.9 Am3/min, 1.8% for 18.9 Am3/min, and 0.7% for
16.9 Am3/min. The average aerosol "d " was 1.2 ymA with a
geometric standard deviation of 2.2.
Particle penetration versus particle size were plotted
for the data obtained from the Electro-Tube and appear in
43
-------
Figures 24, 25, and 26. The data have been plotted together
according to the gas flow rate in the device. The penetrations
calculated from the diffusion battery data were converted to
penetrations corresponding to aerodynamic diameter using both
3 g/cm3 and the 4.1 g/cm3 densities, and are also shown in Figures
24, 25, and 26. The lower density results in 5 to 151 lower
particle penetration for the three sets of runs shown.
MATHEMATICAL MODEL
The APS Electro-Tube is essentially a wetted wall ESP.
The mathematical model presented by Gooch and Francis (1975)
for predicting the performance of an ESP is applicable
ECONOMICS
No data were available for determining the annualized
costs associated with operating the APS Electro-Tube. It is
expected that the costs would be comparable to those of a
conventional wetted wall ESP.
44
-------
0.1
0.05
z
o
cc.
V-
o
< 0.01
E-^
Z
w
ex
pj 0.005
0.001
DIFFUSION
BATTERY
RUN 19
_L
RUN 16 —
0.05 0.1 0.5 1.0 2.0 3.0
AERODYNAMIC PARTICLE DIAMETER, d , pmA
Figure 24. Penetration versus aerodynamic
particle diameter for low gas
flow, Runs 15, 16, 19
0.3
o
H
o
E-
<
C£
H
U
z
ta
ex
0.1
0.05
0.01
0.005
0.001
0.05
DIFFUSION
BATTERY
RUN 17
I I
0.1
1.0
2.0 3.0
AERODYNAMIC PARTICLE DIAMETER, dpa>ymA
Figure 25. Penetration versus aerodynamic
particle diameter for medium gas
flow, Runs 7, 10, 12, 13, 14.
-------
1.0
0.5
o
0.1
OS
U.
z:
o
E-
W
2
w
D-
W
J
CJ
I— I
H
Oi
<
D,
0.05
0.01
0.005
0.001
DIFFUSION
BATTERY
RUN 18
I
0.05 0.1
RUN 5
RUN 18
I f
0.5
1.0
2.0 3.0
AERODYNAMIC PARTICLE DIAMETER, d ,ymA
pa
Figure 26.
Penetration versus aerodynamic
particle diameter for high gas
flow, Runs 3, 4, 5, 18.
46
-------
SECTION 6
TRW CHARGED DROPLET
SOURCE AND CONTROL SYSTEM
The charged droplet scrubber was developed by TRW Systems.
Instead of charging the particles as in the case of APS elec-
trostatic scrubbers, the TRW charged droplet scrubber charges
the water drops. The water flows out of small diameter tubes
which also act as electrodes. The water is atomized as it
jets from the tubes. Particle collection of this scrubber
results from inertial impaction and the electrostatic force
that exists between the particle and the water drop.
The TRW Charged Droplet Scrubber was a 3-stage 680 Am3/min
(24,000 ACFM) pilot unit controlling emissions from the exhaust
of a coke oven. Figure 27 is a schematic diagram of the control
system. The emissions were a side stream from a main stack
and enter a 1.5 meter inlet duct at 215°C. The gases were
then quenched by a water spray system at the rate of 17.0 to
22.7 liters per minute to lower the gas temperature to 121°C
or 93°C, respectively.
The gas then entered an 8.4 kW (11.2 HP) forced draft fan
which had a capacity of 680 actual m3/min (24,000 ACFM) at a
AP of 10.1 cm W.C. The gases then entered through a duct with
straightening vanes leading into the scrubber. The estimated
pressure drop of the entire scrubber system was 10.1 cm W.C.
and the pressure drop for the electrical system was 1.3 cm W.C.
From the scrubber system which had three sections plus a high
voltage area the gases exited through the 0.91 meter exit stack.
The temperature of the gas at the stack exit was 54-66°C.
The particles from the coke oven exhaust were sticky tar,
heavy hydrocarbons, light oil and complex compounds. Cooling
47
-------
QUENCH
SPRAY
U
0 /
0 I
o V
MAIN
STACK
FORCED
DRAFT FAN
TREATED
GAS OUT
3rd STAGE
2nd STAGE
1st STAGE
HIGH VOLTAGE
AREA
STRAIGHTENING
VANE SECTION
(FOUR VANES)
<=
ELECTRODE
WATER WASH
PROBE
ELECTRODE
COOLING
"- AIR
SLUDGE
Figure 27. TRW Charged Droplet Scrubber schematic.
-------
air was used on the high voltage electrode which had to be
accounted for when analyzing the exit gases.
Scrubber performance is independent of water conductivity
in the range of 50 to 1,500 umhos/cm. Domestic water which
was used during the testing period has a conductivity within
this range.
TEST METHOD
The performance characteristics of the TRW Charged Droplet
Scrubber were determined by measuring the particle size distri-
bution and mass loading of the inlet and outlet gas samples
simultaneously.
For the tests performed in October and November 1975,
modified EPA type sampling trains with in-stack and ex-stack
University of Washington Mark III (U.W.) cascade impactors
were used for particle measurements above 0.3 ymA in diameter.
During an impactor run, several inlet and outlet fine
particle size measurements were taken alternately with the
portable screen diffusion battery. Since the system remained fairly
constant during each run, alternate inlet and outlet SDB
measurements were considered to approximate simultaneous
sampling.
Impactor blank runs were performed periodically to assure
that the greased aluminum substrates did not react with the
stack gases. A blank impactor run consists of an impactor
preceded by two glass fiber filters run at identical sample
conditions as the actual sampling runs.
The inlet sample port was located 0.76 meters from the
fan outlet in a flow area reducing section with four
straightening vanes. The water quench spray section was
near the fan inlet. When a velocity traverse was performed
on the inlet, eddy mixing and condensation in the pitot tube
was evident since some negative velocity heads were measured.
The best one point sampling location was taken during the
impactor runs. The eddy mixing indicates that the average
gas velocity and the gas flow rate (volume/time) are question-
able based on the inlet traverse. The outlet port was located
49
-------
three duct diameters downstream of the nearest disturbance
and one duct diameter upstream of the stack outlet. Velocity
traverses of the outlet revealed fully developed flow profiles.
OPERATING CONDITIONS
The operating conditions of the TRW Charged Droplet
Scrubber for the testing period are shown in Table 9.
TABLE 9. OPERATING CONDITIONS
Condition
Temperature, °C
Velocity, m/sec
Am3/min
ACFM
DNm3/min @ 0°C
DSCFM @ 21°C
Vol 1 H20 vapor
Static Pressure
Inlet
Prior to
Quench Sprays
210
2-9
Inlet (1) @
Straightening
Vanes Section
107-121
6.5-7.4
637.0-727.3
22,500-25,690
383.2-437.6
14,580-16,650
7-18
-9.3 cm W.C. +2.5 cm W.C.
Outlet (2)
@ Outlet
Stack
49-63
7.0-7.9
274.3-311.7
9,690-11,010
200.9-228.3
7,650-8,690
7-15
+1.2 cm W.C.
Water quench spray
Electrode water rate
Electrode voltage
Area for gas flow in scrubber
Gas velocity in scrubber (3)
Electrode probe cooling air(4)
Current
17.0-22.7 £/min
45.4-60.6 £/min
31,000-38,000 volts
7.36 m2
0.62-.71 m/s
11.7 std mVmin @ 21°C
205 mA @ 37 kV, 152 mA @ 32.7 kV
(1) Inlet velocity traverse revealed eddy flow patterns and
condensation, e.g., minus velocity heads at some traverse
points. This is due to the inlet sample location being in
a flow area reduction section with straightening vanes, water
condensation in the traverse line due to the quench sprays,
and close location to the blower fan outlet. Therefore the
data presented for flow are suspect.
(2) More reliable flow data than inlet data.
(3) Based on outlet AmVmin.
(4) Electrode probe cooling air accounted for as dilution air.
50
-------
DATA
Cascade Impactor Data
A total of twenty simultaneous inlet and outlet impactor
sample runs were performed on the TRW Charged Droplet Scrubber
and are summarized in Table 10. The runs were performed for
the series of test conditions that TRW had established for
their own tests. Table 11 provides a cross reference between
the runs and the TRW test matrix.
TABLE 10. TEST RUN SUMMARY
Date
10/14/75
10/16/75
10/17/75
10/21/75
10/21/75
10/22/75
10/24/75
11/4/75
11/5/75
11/5/75
11/6/75
11/6/75
11/7/75
Run
No.
1
4
6
8
9
10(1)
13(1)
14
15
16
17
18
19
Scrubber
Velocity
(m/s)(2)
0.62
0.71
0.62
0.71
0.71
0.71
0.62
0.71
0.71
0.71
0.62
0.62
0.71
Electrode
Voltage
(kV)
31
35
35
31
33
38
33
38
38
38
38
38
33
Mass Lo
(mg/D
Inlet
232
163
188
247
132
166
953
201
419
182
508
169
217
adings
Nm3)
Outlet
35.8
61.0
11.0
29.2
41.5
84.5
56.6
23.1
35.5
22.2
23.1
21.9
25.5
Pt
%
15.4
37.5
5.8
11.8
31.4
50.7
5.9
11.5
8.5
12.2
4.5
13.0
11.7
Runs 2, 3, 5, 7, 11, 12, 20 were not included because of
field sampling problems with the weighing balance or
sampling equipment train.
(1) For all runs the inlet loading was measured after
the quench sprays. For runs 10 and 13 only the
inlet was also sampled before the water quench
sprays and are noted as runs 10P and 13P.
(2) Scrubber velocity was based on the outlet Am3/min.
51
-------
TABLF. 11. CROSS-REFERENCE TO TRW TEST MATRIX
RUN NO. IN
THIS REPORT
1
4
6
8
9
10
13
14
15
16
17
18
19
RUN NO. IN
TRW TEST MATRIX
TRW CONTROL
VARIABLES
12
13
14
10
10
15
11
2
7
7
8
8
4
A0
AI
AI
A0
A0
A!
A0
AI
AI
AI
AI
AI
A0
B0
Bl
BO
BI
BI
B!
BO
Bi
BI
BI
BO
B0
Bi
C
C
C
C
C
C
C
C
C
C
C
C
C
0
1
1
0
0
0
i
0
i
j
!
1
!
D
D
D
D
D
D
D
D
D
D
D
D
D
l
0
!
0
0
1
0
o
j
j
0
0
o
where the control variables were:
Variable
Electrode Voltage
Gas Velocity
In Scrubber
Electrode Water
Flowrate
Pre-Cooling Water
Flowrate (Water
Quench Spray)
Symbol
B
D
Low Level
A0(Runs 1-8,
31,000 volts)
(Runs 9-16,
33,000 volts)
High Level
tl(Runs 1-8,
35,000 volts)
(Runs 9-16,
38,000 volts)
B0(0.62 m/sec) Bj(0.71 m/sec)
C0(45.4 Z/min) ^(60.6 A/min)
D0(17.0 JL/min) 0^22.7 Z/min)
52
-------
Particle concentration, particle size, sampled volumes, and
size distributions are tabulated in the Appendix A, Tables A-l
through A-13. Size distributions for the impactor runs are given
in Figures A-l through A-13. A summary of the inlet and outlet
size distributions tests is given in Table 12.
Average sample times for the inlet were ten to twenty minutes
depending on the mass loading, while the outlet sample times averaged
approximately 45 minutes.
The sample data obtained were fairly consistent depending on
operation of the scrubber. No data were obtained for runs 2, 3,
5, 7, 11, 12, and 20 because of field sampling problems with
weighing balance and sampling train. Other runs were of ques-
tionable value because of impactor problems, sampling train pro-
blems, entrainment problems, and upset conditions (heavy loading)
of the TRW Charged Droplet Scrubber.
Run 1 experienced inlet impactor hole plugging on the 5th and
6th stages of up to 50%. Impactor hole plugging would cause the
stage cut diameter to be lower than it would be.
Run 4 experienced heavy loading on the outlet impactor for
15 to 40 minutes total run length. Since the heavy loading was
at the end of the run it would cause the penetration to be higher.
For Run 9 the outlet impactor's 6th and 7th stages were very
wet indicating condensation of water vapor or hydrocarbon materials.
Run 10 experienced heavy loading conditions during the sam-
pling period. When heavy loading conditions were not simultaneous
on inlet and outlet impactor runs faulty penetration data are ob-
tained.
Run 15 is questionable because the inlet 3rd and 5th impac-
tor stages had holes plugged 20-30%. Also the inlet sampling was
during heavy loading 1 out of 18 minutes and mild loading (indi-
cating a loading above normal but still not heavy) 3 out 18
minutes while the outlet loading was heavy 9 of 30 minutes and
mild 7 of 30 minutes.
Run 18 on the outlet sampling had a heavy loading for 8
of 40 minutes at the end of run. Noting that the inlet sampling
ran only 24 mintues and was shut off when the heavy loading occurred
on the outlet, one would expect higher penetration than actual.
53
-------
TABLE 12. INLET AND OUTLET SIZE DISTRIBUTION SUMMARY
Run
No.
1
4
6
8
9
10
13
14
15
16
17
18
19
dpg,ymA
1.37
0.89
0.96
0.59
0.41
2.10
1.02
1.27
1.35
0.67
1.55
1.15
1.19
Inlet
ag
2.7
2.0
2.0
2.1
2.6
4.3
1.9
4.2
2.1
2.5
2.2
8.0
2.8
dpg,ymA
0.57
0.69
0.96
0.41
0.61
0.95
1.13
1.00
1.40
0.83
1.06
1.25
0.98
Outlet
°g
2.6
1.9
2.0
2.7
1.8
2.1
2.0
2.2
2.5
2.3
2.2
2.7
2.5
54
-------
Diffusion Battery Data
Diffusion battery data were taken on December 4 and 5, 1975.
Four inlet runs were attempted. Runs 1 and 4 inlets yielded
the only usable data. Runs 2 and 3 had to be aborted due to
cascade impactor plugging, upset conditions, and blown fuses
on the sampling train. Sampling on the outlet yielded no data
for the following reasons: Condensation in plastic tubing on
the Gardner CNC caused fogging on the upper lens of the Gardner
and this problem could not be alleviated.
Figure A-14 in the Appendix contains the inlet size distri-
butions for the two diffusion battery runs that were usable.
Since no outlet data were obtained penetrations could not be
calculated.
PARTICLE PENETRATION
Particle penetration versus particle size was calculated
for the data obtained from the TRW Charged Droplet Scrubber.
Figures 28 through 31 represent penetration curves obtained
from cascade impactor data for four groups of data sets for
low to high electrode voltage and for low to high gas velocity.
Experimental test results on the TRW Charged Droplet
Scrubber showed overall collection efficiencies ranging from
94.1% for low gas velocity and low electrode voltage and
94.91 for low gas velocity and high electrode voltage. These
results are based on runs that did not have any problems during
the testing period such as heavy loading conditions, entrain-
ment, or impactor hole plugging.
The penetration was calculated by fitting the inlet and
and outlet particle size data and by the method described in
a previous section.
ECONOMICS AND OPERATION PROBLEMS
The Charged Droplet Scrubber tested was a pilot unit.
Numberous scrubber upset conditions (heavy loading) were en-
countered during the sampling period. A high efficiency en-
trainment separator should be incorporated into the scrubber
system because heavy entrainment from the scrubber was visible.
55
-------
1.0
AERODYNAMIC DIAMETER, d , ymA
pa
•Impactor 5th § 6th stage holes
plugged up to 50%
Figure 28. Particle penetration versus aerodynamic
diameter for low electrode voltage and
low gas flow rate.
1.0
0.5
•H
<->
-
-
0.1
0.05 hSg
0.01
0.2 0.5 1.0 5.0
AERODYNAMIC DIAMETER, d , ymA
pa
*4 - Heavy loading on outlet 15 of 40 minutes @ end of run.
*10 - Heavy loading conditions while sampling.
*15 - 3rd to 5th impactor stages 20-30% plugged.
Inlet 1 of 18 minutes heavy loading, 3 of 18 mild loading.
Outlet 9 of 30 minutes heavy loading, 7 of 30 mild loading.
Figure 29. Particle penetration versus aerodynamic
diameter for high electrode voltage and
high gas flow rate
-------
-
3
-
-
^
-
I
7
£
0.02
0.1
0.05 g
0.2
0.5
AERODYNAMIC DIAMETER, d , umA
pa
*18 - Heavy loading on outlet sample 8 of 40 min.
till end of run - then very heavy loading.
1.0
0.03
AERODYNAMIC DIAMETER, d umA
pa*
*9 - Impactor 6th and 7th stages very wet
Figure 30. Particle penetration versus aerodynamic
diameter for high electrode voltage and
low gas flowrate
Figure 31. Particle penetration versus aerodynamic
diameter for low electrode voltage and
high gas flow rate.
-------
Cost data for the TRW Charged Droplet Scrubber were not
provided by the manufacturer.
The theoretical power consumption for the electrical
power section of the TRW Charged Droplet Scrubber was 25.9
W/(m3/min) at 37 ^y operation and 17.0 W/(m3/min) at 32.7 kV
operation. Power consumption for the fan section was 33.3
W/(m3/min) based on a fan and motor efficiency of 50% and
a total system pressure drop of 10.1 cm W.C.
A venturi scrubber with equivalent overall collection
efficiency and cut diameter would have a pressure drop of
approximately 100 cm W.C. Based on a fan and motor efficiency
of 501, the power consumption would be 354 W/cm3/min.
Based on 8,000 hours operation per year and a cost of 3
-------
SECTION 7
PERFORMANCE TEST METHOD
The method of approach to the program objectives involved
a number of experimental determinations to obtain collection
efficiency data, the aquisition of information on system
characteristics and behavior, and computations which utilized
the performance data and mathematical models. Over the course
of the program the methods and apparatus used were generally
improved and were modified to suit each specific test situa-
tion but the main features were similar and will be described
here.
The most important experimental measurements were those
regarding particle size and concentration. Cascade impactors
were used for particle measurements above 0.3 ymA. The Air
Pollution Technology portable screen diffusion battery (A.P.T.-
S D B ) was used for particle measurements from 0.01 ym to
0.1 ym (actual). The apparatus and methods used are outlined
below.
1. Gas velocity distribution and parameters had to be
measured at the inlet and outlet of the scrubber in order to
define the following:
a. Conditions for isokinetic sampling.
b. Particle concentration per unit volume of dry gas, which
is a consistent basis for comparing inlet with outlet
in the computation of efficiency.
c. Gas flow rate.
d. Amount of liquid entrainment in the outlet.
59
-------
Methods to measure these parameters are tabulated below:
Parameter
Equipment
Method
Gas velocity
and flow rate
Gas temperature
Humidity
Pressure
Standard pitot tube or cali-
brated type "S" pitot tube;
differential pressure gauge.
Calibrated thermocouple or
mercury filled glass-bulb
thermometer.
Thermometers.
Inclined water manometer
or a pressure gauge.
EPA Method 1;
EPA Method 2.
Wet and dry bulb
temperature mea-
surement on a
flowing sample
withdrawn from
the duct.
Measured by means
of a static pres-
sure tube inser-
ted in the duct.
2. The most essential part of the scrubber performance
tests is the determination of particle size distribution and
concentration (loading) in the inlet and outlet of the scrubber.
For accurate determination of particle size distribution, a
collection mechanism that collects particles and causes neither
formation nor breakup of aggregates is necessary. Cascade
impactors come close to meeting these requirements.
In a cascade impactor, particles are classified by inertial
impaction according to their mass. The larger ones are collected
on the plate opposite the first stage and the smallest on the
plate opposite the last stage. A.P.T. uses Brink, Andersen,
University of Washington Mark III, and a cascade impactor of
A.P.T.'s own design for particle size fractionation. These
impactors (except Brink) classify particles into seven size
groups and are capable of sizing particles down to about 0.1 ym
diameter (actual). All impactors were calibrated in the
laboratory according to EPA guidelines, see Calvert et al. (1976c)
and Harris (1977) on calibration method.
60
-------
In order to minimize probe losses all tests were made with
the impactors in the duct and with the inlet nozzles appro-
priately sized to give isokinetic sampling. A modified EPA
Method 5 train was used to monitor the sample gas flow rate.
Figure 32 shows the sampling train arrangement.
In some tests, a pre-cutter was used to remove either the
heavy particle loading from inlet samples or the entrained
liquid from outlet samples. A round jet impactor with about
8 ymA cut diameter was found to have good characteristics and
was adopted for use for both inlet and outlet sampling. The
impactors were either given time to reach the duct gas tempera-
ture or heated to prevent condensation.
To increase the weighing accuracy, light weight substrates
were used on the collection plates. Generally, either greased
aluminum foil or a glass fiber filter paper substrate was
used. Impactor substrates and back-up filters were weighed
with an analytical balance to the nearest tenth milligram (10~"g)
Particle size distribution and loading measurements were
conducted simultaneously at the scrubber inlet and outlet. The
method minimizes the effects of particle size distribution
changes caused by fluctuations in the operation parameters.
Since the program objective was to investigate scrubber perfor-
mance on fine particles, the sampler was held at one location
in the duct for the duration of each sampling run. This is an
adequate technique for obtaining good samples of particles
smaller than a few microns in diameter because they are
generally well distributed across the duct.
Blank impactor runs were performed periodically to assure
that the greased aluminum substrates did not react with the
stack gases or lose weight. A blank impactor run consists of
an impactor preceded by two glass fiber filters and run at
identical sample conditions as the actual sampling runs.
3. In-stack filter samples were also taken to obtain
total particulate loadings and overall collection efficiencies
of the system. The sampling train arrangement was the same
as the cascade impactor train except the cascade impactor was
61
-------
tsj
PRECUTTER
AND
NOZZLE
LJ
THE I
HI: ATE D
CASCADE
IMPACTOR
ORIFICE
METER
7
/
•>
STAC
, ,
L
K
WALL
IMPINGER TRAIN
|
MANOMETER
DRY GAS
METER
VACUUM
PUMP
SILICA
GEL
DRYER
Figure 32. Modified EPA sampling train with in-stack cascade impactor.
-------
replaced with a filter.
4. Inertial impaction devices (cascade impactors) are
normally insufficient for measurement of particulates less
than 0.3 ym (actual diameter). Fractionation of these par-
ticles is best accomplished by diffusional collection devices,
or diffusion batteries, usually consisting of closely spaced
parallel plates or long, thin tubes. Large quantities of
pumps, dilution apparatus, and other battery related equipment
are bulky and prove to be cumbersome in field use. For this
and other reasons, Air Pollution Technology, Inc. developed a
portable screen diffusion battery which is lighter and more
mobile than previous devices.
The Screen Diffusion Battery utilizes a series of layered
screens intermittently separated for sampling purposes (Figure
33). Size fractionation by the diffusion battery is detected
by measurement of overall particle concentrations of the gas
stream into and out of a known number of screens using a
condensation nuclei counter (CNC). Concentrated aerosol samples
are diluted until compatible with the CNC (~106 particles/cm3).
Screen penetration data are then analyzed to determine size
distribution and cumulative mass loading of the particulates
in the stream. If desired, cascade impactor and diffusion
battery analyses can be combined and an overall characteriza-
tion of the particulate size distribution (and scrubber pene-
tration) obtained.
Fine particle size measurements with the diffusion battery
were not taken simultaneously at the inlet and outlet of the
scrubber system. During an impactor run, several inlet and
outlet measurements were taken alternately with the S D B
Since the system remained fairly constant during each run,
alternate inlet and outlet S.D.B. measurements were considered
to approximate simultaneous sampling.
Each S D B run consisted of a continuous series of CNC
readings. Normally, CNC counts were taken at each diffusion
battery stage in order of increasing number of screens and
then the process was repeated until three to four sets of
63
-------
r
STACK
GAS
1st DILUTION
FLASK WITH CHARGE
NEUTRALIZERS
CASCADE
ffl
FILTER
O
ROTAMETER
VACUUM
PUMP
IMPACTOR
i
I
•w
(
^
^
1
w
^
' ROTAMETERS
••
?
x
]
-— .
FI
•\
^ «^ «^
.»
T P
~ 1 1
1
LTER A
^ urnw V O
»«-
>*
•*«•
DIFFUSION
BATTERY
METER 1 ()
^ ^ 1 T
2nd
DILUTION
FLASK ^
TO
CNC
PUMP
VACUUM
DESSICANT
Figure 33. Schematic diagram of diffusion battery system.
-------
readings were obtained. Continuous monitoring of flow,
temperatures, and pressures enabled steady operation of the
diffusion battery. Conditions in the duct (pressure, gas
velocity, temperature, and water vapor content) were obtained
during the impactor tests.
The size of particles entering the diffusion battery was
limited by using a cascade impactor pre-cutter on the in-stack
end of the probe. Isokinetic sampling was not maintained
because the particles to be measured were too small to be
segregated by inertial effects from bends in the gas stream.
The sample stream entering the diffusion battery was immediately
diluted with heated, dried, filtered air to control condensation.
Two Polonium 210 charge neutralizers were inserted into the
flask to eliminate electrostatic effects. A portion of the
resultant aerosol was directed through the diffusion battery
and the outlet diluted to a concentration measurable by the
Gardner CNC. The aerosol from the second dilution flask was
sampled with the Gardner CNC. The excess aerosol was exhausted
through the vacuum pump. The excess aerosol from the first
dilution was passed through an absolute filter and pumped
to the atmosphere. The Gardner CNC was calibrated daily against
a standard B.G.I. Pollak, Model P, CNC and found to read con-
sistently 33% lower than the Pollak CNC for the concentration
range used in the testing.
ERROR ANALYSIS
Sample Bias
It is important to note that the program objective is to
investigate scrubber performance on fine particles and, con-
sequently, it is not necessary that the methods used be
accurate for large particles. This makes the sampling
simpler in the following ways:
1. Isokinetic conditions are not important for fine
particles. For example, the error caused by sampling 4 umA
particles at a velocity 50% higher or lower than the gas
stream velocity would only be about 2 or 3% of the concentration.
65
-------
2. The fine particles will be well distributed in the
gas stream, except in cases where streams with different
particle concentrations have not had time to mix, so single
point sampling is generally sufficient. To illustrate, we
may note that the Stokes stopping distance of a 3 umA particle
with an initial velocity of 15 m/sec (50 ft/sec) is about
0.04 cm (0.016") and for a 1 umA diameter particle it is one
ninth of that. Since the stopping distance is the maximum a
particle can be displaced from a gas stream line by going
around a right angle turn, it is obvious that fine particle
distribution in the gas stream will be negligibly affected
by flow direction changes.
3. The effect of a pre-cutter on the size resolution of
a cascade impactor is not significant in the size range of
interest, so long as the pre-cutter has a cut diameter larger
than several microns.
Diffusion Battery
The Screen Diffusion Battery was calibrated in Air Pollu-
tion Technology's (A.P.T) small particle laboratory. An aerosol
of known size distribution was generated and passed through
the diffusion battery. The total number concentration was
measured with a condensation nuclei counter at the battery
inlet and outlet of each S D B stage. The penetration of
particles (percent) was then calculated and plotted against
solidity factors on semi-logarithmic paper. The experiment
was repeated with the same aerosol until a smoothed average
curve relating number penetration to solidity factor was
obtained. From the smoothed curve, a correction factor for
the theoretical diffusion battery performance was determined.
The scatter of data points about the smoothed (fitted)
calibration curve represents the experimental error in the
penetration measurement. This measurement error included
meter reading error, accuracy of the CNC, etc. The measure-
ment error was defined in terms of relative error, or the
deviation from the averaged penetration value divided by the
averaged value.
66
-------
This procedure was repeated on other aerosols of known
size distribution. The maximum relative error was then deter-
mined from these experiments for each solidity factor. The
maximum relative error the Screen Diffision Battery determined
by this method is 10.41 for solidity factors of 13, 26, and 40.
Cascade Impactors
Cascade impactors were used as the principal means of
obtaining information about the inlet-outlet size distributions.
It was important to understand the sources of error and how the
error can be minimized.
The procedural errors include the accuracy of the weighing
of the deposits, reading of the test data such as temperature,
gas volume, time, and pressures. The errors from the impactor
design and construction include wall losses, accuracy and
precision in construction of critical components, and particle
re-entrainment from the collection surface.
Some of the design and construction limitations can be
reduced by procedures such as recovering the wall losses and
by sampling at certain flow rates and times to reduce re-entrain-
ment errors and by calibration of the impactor. The experi-
mental data obtained with commercial impactors were reported
by Smith et al. (1974). Smith et al (1974) reported that all
impactors tested had appreciable wall losses for particle
diameters above 10 microns. This error can be reduced by
brushing the material from the wall onto appropriate collec-
tion disks. The flow velocity through the impactor jets should
not be above 65 m/sec to be absolutely certain of avoiding
re-entrainment of particles from the collection substrate.
The extent of re-entrainment will depend on the properties of
the material and the amount of deposit. Lundgren (1967)
reported that re-entrainment increased as the collection sur-
face became coated with particles. However, Rao (1975)
reported that collection efficiency increased with increased
particle load. When the particle weight is over 10 mg, part
of the deposit may break away from the surface and migrate
within the impactor. The light-weight deposit places
67
-------
importance on accurate weighing. The analysis of impactor
errors was limited to the weighing error and in the calculation
of collection efficiency error. The effects of weighing errors
on the results of impactor tests have been analyzed by Sparks
(1971). An analysis of the weighing error using three different
estimations was reported by Fegley et al. (1975). The results
indicate that, when the weight of sample per stage is less
than 1 mg when weighed with a balance with a precision of
0.05 mg, the error in the fractional mass will be greater than
10%.
68
-------
SECTION 8
DATA REDUCTION AND COMPUTATION METHOD
CASCADE IMPACTOR DATA ANALYSIS
In a cascade impactor particles are classified by inertial
impaction according to their mass. The larger ones are
collected on the plate opposite the first stage and the
smallest on the plate opposite the last stage.
Once the stage "catches" have been measured, usually by
weighing particle collection foils or papers, the data analysis
is relatively simple. Generally the objective is to make a
plot of particle diameter versus mass percent oversize or
undersize and to represent the size distribution in terms of
log-normal distribution parameters if possible. Thus, it is
necessary to do the following:
1. Add all of the stage and filter collection weights
to get the total particle mass collected.
2. Compute either:
a. Cumulative percent collected as the gas flows
through succeeding stages. This is "percent
oversize".
b. Cumulative percent penetrating as the gas flows
through succeeding stages. This is "percent
undersize".
3. Compute the cut diameters for the impactor stages,
taking into account gas viscosity (or temperature)
and gas sampling flow rate. The equation is:
d = K°-5 /9 ^r di \°'5
dpac KP5o| G 3 1 (8)
u. x 10"8
69
-------
where d = impactor stage cut diameter, ymA
p ciC
yG = gas viscosity, poise
d. = jet diameter, cm
u- = jet velocity, cm/s
K = inertial impaction cut parameter, K at 50%
P50 efficiency p
The particle diameter used is called "aerodynamic diam-
eter" and it has the unit of "aerodynamic microns", ymA.
This is the effective diameter for particle separation by
inertial impaction and it takes into account the effects of
particle density and particle "slip" between gas molecules.
It is related to the actual physical size of the particle
by the following equation:
(-• *)'
dpa ' dnK C'l2 C 9)
where d = aerodynamic particle diameter, ymA
pa
d = actual particle diameter, ym
p = particle density, g/cm3
Cf = Cunningham slip correction factor, dimensionless
At room temperature for air the Cunningham slip correction
factor, C', is given by:
C' = 1 + °-i65 (10)
P
Figure 34 is a plot of aerodynamic particle diameter
versus particle diameter for various particle densities.
If the particle distribution follows the log-normal
law, a straight line will result on log-probability paper.
The 50% value of particle diameter is the mass median
diameter "d " and the geometric standard deviation "a " is
r o
given by:
84.31 value of d^
CT =
g 50% value of d
70
(11)
-------
50
w
H
0
w
u
0.1
1.0
PARTICLE DIAMETER, ym
Figure 34. Particle diameter versus particle
aerodynamic diameter.
10
71
-------
OVERALL PARTICLE PENETRATION
Overall particle penetration is defined as
C
= 7^- / Pt, cT C
(12)
*"• •&
mass concentration out
mass concentration in
where Pt = overall particle penetration, fraction or percent
Ptd = Penetration for particles with
diameter d , fraction
C = total particle weight, g
pt
C = mass of particles, g
The overall particle penetration can be computed using
the data from a simultaneous inlet and outlet cascade
impactor or filter run.
PARTICLE PENETRATION AS A FUNCTION OF PARTICLE DIAMETER
The particle penetration for particle diameter, d , or
grade penetration curve is given by
f (dloutlet
Pt, = „, P .
d f (d )inlet
d(y
d
outlet
(13)
inlet
where Pt^ = particle penetration for particle diameter,
d •, fraction
f(d ) = particle frequency distribution
dC
= the slope of the cumulative mass versus
P particle diameter curve at d , dimensionless
Particle penetration as a function of particle size is
computed from inlet and outlet particle size distributions
and concentration data. The major steps involved in the
computation are as follow:
1. Reduce cascade impactor data to the form of cumulative
particle mass concentration for each impactor cut diameter.
2. Determine the slopes of the cumulative mass distri-
bution curves at several values of particle diameter for
72
-------
both the inlet and outlet and then compute penetration at
each particle diameter.
There are several techniques to determine the slope of
the cumulative mass distribution curve. Some of the techniques
are discussed below.
1. Graphical technique - Cumulative mass concentration
versus aerodynamic particle diameter data may be fitted with
a curve by eyeball method. The slopes of the curve at various
values of particle diameter are then measured graphically.
2. Curve fitting - Curve fitting to the data points and
the measurement of curve slopes by eye involves subjective
judgment. To eliminate the judgment errors, it is possible to
fit the data with a mathematical function and then evaluate
the slope analytically. We have tried fitting the cumulative
mass curves with polynomial functions, log-normal distribution
functions, and the Weibull distribution
Polynomial Curve Fit
The principle of least squares was used to fit a function
of the following form to the data points
Wcum = a° + ai dpa + a* dpa + ••" + an dp£ ^
where wcum = cumulative mass concentration, g/cm3
d = aerodynamic particle diameter, ymA
"ao, ai, 82, .... an" are constants to be determined. A second
degree polynomial function will normally fit the data points
closely.
Curve smoothing is also required when using a polynomial
function because the resulting curve oscillates. The inlet
and outlet polynomial curves do not oscillate in the same
manner. Therefore, without curve smoothing, the calculated
scrubber penetration curve will also oscillate.
Log-normal Distribution
If the inlet and outlet size distributions are nearly
log-normal, then a purely mathematical particle penetration is
used. The mathematical log-normal penetration is based on the
73
-------
following:
The distribution of particle sizes is:
P(x) = — L-p /*Vt2/23t (15)
W2-i
W
where P = **-, the cumulative mass fraction of sizes
t
smaller than "d ", and
pa
In ag
The derivative of the distribution function is
d P = d P d" x = exp (-x2/2)
"
thus,
dpa d * d dpa C21T)-* dpa in ag
d W Wt
d d (2ir) 2 d In a
pa pa g
Using equation (5), the penetration is then
(W.) (In a_). /x2 . x
t^out g in
_. . 2 , \
g in PYT) / in - _ out 1
' WI7 (ln "Pout ' ( 2 ^
Weibull Distribution
The Weibull distribution (Lipson § Sheth, 1973) offers 2 advan
tages over the log-probability distribution. The first is that
it has three parameters rather than two. The second is that
the cumulative distribution function (CDF) is explicit and
does not have to be approximated by multi- termed polynomials.
74
-------
Cumulative Distribution Function-
CDF = 1 - exp
d - d
P
0 - d
po
(2Q)
where d = particle diameter
minimum particle diameter
characteristic diameter
b = Weibull slope
po
0
The CDF has the property that:
CDF
d ) = 0.632
(21)
The median particle diameter occurs when the CDF = 0.5, so:
Pg
- V (ln 2)
1/b
Linear Transformation -
Transformation to a linear form;
(22)
y = A + B x
(23)
requires that:
y = In
therefore,
.-CDF,
- V
(24)
(25)
A = -b In (6 -d )
B = b
(26)
(27)
75
-------
and,
0 = dpo + exp - I (28)
b = B (29)
Least Squares Curve Fit -
The minimum particle diameter, d , is that which results
in the highest linear correlation coefficient based on the
above linear transformation, when a least squares linear
regression is performed on the size distribution data. Note
that,
0 £ d < smallest diameter found in the distribution
Density Function -
The Weibull density function is the derivative of the CDF:
ffd
c
d-d
vb-1
P
-d
.b"
(30)
Penetration -
The penetration is the ratio of the cumulative mass loading
distribution derivatives,
where CT = total mass concentration or loading
"out" = refers to outlet particle size distribution
"in" = refers to inlet particle size distribution
and f(d ) = is defined by equation (30J
Minimum Particle Diameter, Physical Interpretation -
"d " is the smallest diameter of the total distribution.
po
A value other than zero means that the data indicate that there
is a minimum particle size. This is physically reasonable
76
-------
because of the particle formation mechanisms and possible
agglomeration and/or particle growth.
Characteristic Diameter, Physical Interpretation -
"0" is analogous to the geometric mean particle diameter
of the log-probability distribution and is therefore an indi-
cation of the "average" size of the particles in the distribu-
tion. The median particle diameter is directly related to
"0" by equation (22).
Wribull Slope, Physical Interpretation -
"b" is analogous to the geometric standard deviation of
the log-probability distribution. It indicates the "spread"
of the size distribution. The larger the Weibull slope, b,
the more uniform (monodisperse) the particle sizes.
SCREEN DIFFUSION BATTERY DATA ANALYSIS
Screen diffusion battery data consist of particle number
concentrations which are obtained after each stage in the diffu-
sion battery. Particle penetrations are calculated from the
ratio of the number concentration taken at a given SDB stage
to the inlet number concentration. The particle size distribu-
tion may then be determined from the penetration data.
Most particle formation processes result in a particle size
distribution which is log normal. Log normal size distributions
are conveniently represented by two parameters: the number median
diameter (d ) and the standard deviation (a ). Typically,
process created aerosols fail log normality only at the extremes
of large and small particles which represent only a small per-
centage of the total particulates.
Calvert et al. (1974) describes a method for converting
log-normal size distributions to overall penetrations using the
relation between particles of a discrete diameter and pene-
tration of those particles through the device.
A log normal particle distribution density function is
defined by:
77
-------
f(dp) =
•y2TT In a
exp
g
- ln
2 ~l
In a
g
(32)
dn
N
dp)
(33)
For particle penetration of the form:
Pt = exp (-AS d B)
(34)
where Pt = penetration fraction of a particle of a given
diameter through S
S = solidity factor, dimensionless
d = particle diameter, ym
A,B = constants established by theory and laboratory
experiment at ion
n = number of particles, dimensionless
Calvert,
-------
Setting:z =
ln dp - ln dpN
(37)
In a
g
and,
-------
exp (-1.29 z ^2 In o ) dz (43)
J
Equation (43) may be evaluated for a number of "a " and
o
normalized values of the cut diameter "d " as shown in Figure
35. Each stage of the SDB has a corresponding cut diameter,
d , for a given flow rate. The cut diameter, d , is the par-
ticle size which is collected at 50$ efficiency. The value of
the cut diameter, d , is determined by substituting 0.5 for
"Pt" in equation (42).
For a given flow rate and cumulative solidity, S, each stage
of the SDB will have a corresponding cut diameter, d , as shown
by the dotted lines in Figure 36. The number median diameter,
d j^,, and standard deviation, a , for the aerosol is determined
by plotting the penetration data for each stage on the dotted
lines. Figure 36 is for a flow rate of 4.7 £/min through the SDB
modules containing 250 mesh screens. The data give a continuous
curve when plotted on one graph.
The value of d N corresponds to the value of "d " for Pt =
0.5. For the aerosol which corresponds to the penetration data
plotted on Figure 3b , the value of d = 0.026 ym.
The standard deviation, a , of the size distribution is
determined by overlaying Figure 36 with the data on Figure 35.
By moving the overlay, with horizontal axes coincident, the
curve may be situated within the family of curves and the
actual "o " may be interpolated by matching the curvatures.
o
Non-log normal data must be handled by a graph stripping
technique outlined by Sinclair (1972) which entails tedious
graphical integration and mathematical conversions. In our
experience the data have fallen sufficiently close to log
normality that the overlay technique is acceptable.
Conversion of number distribution to mass distribution is
necessary in order to put the diffusion battery and cascade
impactor data on the same basis. The method used to make this
80
-------
1.0
oo
-
s
f-
M-t
a:
H
w
I-J
—
w
s
0.1
0.01
iiiiiim
Plot of data before
shifting horizontal axis
0.1
1.0
10.0
50.0
Figure 35. Overall penetration fraction versus "d /d N" with
"a " as a parameter.
o
-------
i. n
c
H
<
C
E-
H
2:
MM
C-
o. ni
SDB|MODULEI'J!
ill
2 h 4 I 5
n.ni
Figure 36,
n. l
PARTICLE DIAMETER, d , ym
Particle penetration through SDK at
0=4.7 i/min.
1.
82
-------
conversion is a graphical integration of the following equation:
X
l —
where N = cumulative number concentration of particles
smaller than "d ", no./cm3
N t = total number concentration of particles, no./cm3
d = particle diameter, ym
d • = a specific particle diameter, ym
mp = mass of particles in the infinitesimal size
range (d" + ddp) , g
M • = cumulative mass concentration of particles
smaller than "d .", g/cm3
The quantity (3m /dNp) is simply the mass per particle of diameter
"dp". The quantity (N x 100/N ) is the number percent of
particles smaller than "d ". Thus, equation (44) can be evaluated
from a plot of mass per particle versus cumulative number percent
of particles, both quantities being evaluated at the same par-
ticle diameter to provide a point on the plot. The total and
cumulative number concentration data are obtained as described
previously.
CUT/POWER RELATIONSHIP
When scrubbers are operated at different pressure drops, it
is very difficult to evaluate and to compare their performance
based only on grade penetration curves. Calvert (1974) has
developed a useful correlation called the cut/power relationship
for this purpose. The cut/power relationship is a plot of the
cut diameter given by the scrubber against pressure drop or power
input, as illustrated in Figure 37. Cut diameter is the particle
diameter whose collection efficiency is 50%. The solid lines
in this graph were calculated theoretically from the design
equations presented in the "Scrubber Handbook" and Calvert (1974).
83
-------
SCRUBBER POWER, .T/kg
If)1*
1.0 -
oc
tu
E-
Note*:
la. Sieve-plate column with foam density of
0.4 g/cm3 and 0.2-in. hole dia. The
number of plates does not affect the re-
lationship much. (Experimental data and
mathematical model.)
1b. Same as la except 0.125-in. hole dia.
2. Packed column with Vin. rings or sad*
dies. Packing depth does not affect the
relationship much. (Experimental data
and mathematical model.)
3a. Fibrous packed bed with 0.012-in.-dia.
fiber—any depth. (Experimental data and
mathematical model.)
3b. Same as 3a except 0.004-in.-dia. fibers
3c. Same as 3a except 0.002-in.-dia. fibers.
4. Gas-atomized spray. (Experimental data
from large Venturis, orifices, and rod-type
units, plus mathematical model.)
Mobile bed with 1 to 3 stages of fluidized
hollow plastic spheres. (Experimental
data from pilot;plant and large-scale power
plant scrubbers.)
I III
n. 5
i.o
10
CAS PHASE PRESSURE DROP, kPa
Figure 37. A.P.T. cut/power plot.
-------
conversion is a graphical integration of the following equation:
Nnt /^P1 dmr, / N^ X 100\
U L I P i I U 1 f A * ~\
i = TOO / dN d -^U (44)
1 P \ V /
where Np = cumulative number concentration of particles
smaller than "d " no./cm3
Npt = total number concentration of particles, no./cm3
d = particle diameter, ym
dpi = a sPecific particle diameter, ym
mp = mass of particles in the infinitesimal size
range (d + ddp) , g
M ^ = cumulative mass concentration of particles
smaller than "d .", g/cm3
The quantity (3m /dNp) is simply the mass per particle of diameter
"dp". The quantity (Np x 100/N ) is the number percent of
particles smaller than "d ". Thus, equation (44) can be evaluated
from a plot of mass per particle versus cumulative number percent
of particles, both quantities being evaluated at the same par-
ticle diameter to provide a point on the plot. The total and
cumulative number concentration data are obtained as described
previously.
CUT/POWER RELATIONSHIP
When scrubbers are operated at different pressure drops, it
is very difficult to evaluate and to compare their performance
based only on grade penetration curves. Calvert (1974) has
developed a useful correlation called the cut/power relationship
for this purpose. The cut/power relationship is a plot of the
cut diameter given by the scrubber against pressure drop or power
input, as illustrated in Figure 37. Cut diameter is the particle
diameter whose collection efficiency is 503. The solid lines
in this graph were calculated theoretically from the design
equations presented in the "Scrubber Handbook" and Calvert (1974).
83
-------
SCRUBBER POWER, .I/kg
oc
Di
UJ
H
UJ
Notn:
la. Sieve-plate column with foam density of
0.4 g/cm3 and 0.2-in. hole dia. The
number of plates does not affect the re-
lationship much. (Experimental data and
mathematical model.
ib. Same as la except 0.125-in. hole dia.
2. Packed column with 1-in. rings or sad-
dles. Packing depth does not affect the
relationship much. (Experimental data
and mathematical model.)
3a. Fibrous packed bed with 0.012 in.-dia.
fiber-any depth. (Experimental data and
mathematical model.)
3h. Same as 3a except 0.004-in.-dia. fibers.
3c. Same as 3a except 0.002-in.-dia. fibers
4. Gas-atomized spray. (Experimental data
from large Venturis, orifices, and rod-type
units, plus mathematical model.
5. Mobile bed with 1 to 3 stages of fluidized
hollow plastic spheres. (Experimental
data from pilot-plant and large-scale power
plant scrubbers.)
0. 5
1. 0
10
GAS PHASE PRESSURE DROP, kPa
Figure 37. A.P.T. cut/power plot.
-------
The cut/power relationship has many useful applications.
It may be used to compare and evaluate scrubbers, to make prelimi-
nary scrubber selections, or to estimate the minimum pressure
drop of a scrubber to attain a required performance level.
Equivalent Cut Diameter
For purposes of comparison it is valuable to have a single
parameter which describes the efficiency of a scrubber. The
cut diameter has proven to be useful for most scrubbers.
Cut diameter is the particle diameter whose collection
efficiency (or penetration) is 50%. It may be read directly
from the grade penetration curve.
In cases where the penetration curve does not reach 501,
the cut diameter cannot be determined directly from the curve,
but an equivalent cut diameter may be determined. The equiva-
lent cut diameter is the particle diameter which would be
required to give the same overall penetration (i.e., over the
entire particle size distribution) if the scrubber had the
same penetration versus pressure drop characteristics as a ven-
turi scrubber. An alternative is to use the equivalent pres-
sure drop that a venturi would need as a criterion for comparison.
The theoretical cut diameter for venturi-type scrubbers
can be calculated using Figure 38. For example, if a = 2.2
and the overall penetration is 0.087, then using Figure 38,
the ratio of cut to mass median diameters is approximately 0.31.
For d = 0.95 ymA and using the above ratio of cut to mass
median diameter the theoretical aerodynamic cut size, "d c"
is 0.29. The cut diameter computed by this method is that
which would be provided by a conventional high energy (venturi
type) scrubber which would give overall penetration equivalent
to the subject scrubber.
85
-------
0.5
CO
0.001
0. 001
Figure 38.
0.01 o.l
dpac/dpg(AERODYNAMIC)
1.0
Overall penetration vs. cut to mass mean particle
diameter ratio for log-normally distributed par-
ticles. (Calvert et al (1972)
-------
The cut/power relationship has many useful applications.
It may be used to compare and evaluate scrubbers, to make prelimi-
nary scrubber selections, or to estimate the minimum pressure
drop of a scrubber to attain a required performance level.
Equivalent Cut Diameter
For purposes of comparison it is valuable to have a single
parameter which describes the efficiency of a scrubber. The
cut diameter has proven to be useful for most scrubbers.
Cut diameter is the particle diameter whose collection
efficiency (or penetration) is 501. It may be read directly
from the grade penetration curve.
In cases where the penetration curve does not reach 50%,
the cut diameter cannot be determined directly from the curve,
but an equivalent cut diameter may be determined. The equiva-
lent cut diameter is the particle diameter which would be
required to give the same overall penetration (i.e., over the
.entire particle size distribution) if the scrubber had the
same penetration versus pressure drop characteristics as a ven-
turi scrubber. An alternative is to use the equivalent pres-
sure drop that a venturi would need as a criterion for comparison.
The theoretical cut diameter for venturi-type scrubbers
can be calculated using Figure 38. For example, if a = 2.2
and the overall penetration is 0.087, then using Figure 38,
the ratio of cut to mass median diameters is approximately 0.31.
For d = 0.95 ymA and using the above ratio of cut to mass
r o
median diameter the theoretical aerodynamic cut size, "d "
is 0.29. The cut diameter computed by this method is that
which would be provided by a conventional high energy (venturi
type) scrubber which would give overall penetration equivalent
to the subject scrubber.
85
-------
0.5
oo
I I—I I I 1111 1—I—I I I 111[
Pt = exp (-Ad 2)
1 pa '
0. 1
Of.
UJ
0. 1
0.001
o =6
8
l i
0.001
Figure 38
0.01 0.1
d /d (AERODYNAMIC)
1.0
Overall penetration vs. cut to mass mean particle
diameter ratio for log-normally distributed par-
ticles. (Calvert et al (1972).
-------
SECTION 9
EFFECTS OF CHARGED PARTICLES ON THE
EXPERIMENTAL PERFORMANCE OF ELECTROSTATIC CONTROL DEVICES
Recent experiments performed at A.P.T. have shown that
cascade impactors may not give accurate size distribution
information when measuring charged particles, Patterson et al.
(1977). Figure 39 shows that the impaction parameter, K , for
a given stage of the U. of W. impactor will decrease with
charged particles. Using the value of "K so" determined for
uncharged particles will overestimate the size of particles
collected.
A more important effect with charged particles which can
be seen from Figure 39 is that the collection efficiency does
not go to zero for low values of the impaction parameter, but
may become constant at. 10 to 20%. Therefore, some of the par-
ticles which should be collected on the lower stages will be
collected on the upper stages. This leads to further over-
estimation of individual particle size.
The mass distribution will likewise be overestimated such
that the larger size fractions will appear to have more mass
than is actually there. This can lead to discrepencies
between theoretical and experimental performance of control
devices. The effect is most noticeable when the inlet dust is
electrically neutral and the outlet particles are charged.
This situation existed during certain performance test runs on
the APS Electrostatic Scrubber and Electro-Tube and the TRW
Charged Droplet Scrubber.
The effect of using cascade impactor calibrations based
on uncharged particles for determining the performance
87
-------
inn
i—r
90
so
A UNCHARGED PARTICLES
O CHARGED PARTICLES
(GROUNDED IMPACTOR)
70
u
rr 50
c
,»j
30
10
i I I I L
OS
.10 .15 .20 .25 .30
IMPACT ION PARAMETER, K
.35
Ficure 39. Impaction characteristics with glass
fiber filter (0.5 pm PSL).
88
-------
characteristics of an electrostatic control device was determined
with the ESP model of Gooch and Francis (1975). The calculation
procedure was as follows:
1. The inlet size distribution was assumed to comprise
uncharged particles with a dpg IN = 10 ymA and a =4.0.
2. An outlet size distribution is computed from the
theoretical curve shown in Figure 41.
3. The outlet size distribution is converted from that
for uncharged particles, curve a, Figure 40 to the recovered
size distribution from a cascade impactor assuming charged
particles, curve b, Figure 40.
4. The collection efficiency is determined as a function
of the particle size using the assumed inlet size distribution
and the recovered size distribution as calculated in number 3
above.
Figure 41 shows the theoretical collection efficiency for
an electrostatic precipitator, the results of these calculations
and some experimental performance data. Therefore the under-
estimation of the collection efficiency for particles greater
than several microns in diameter and the overestimation of the
collection efficiency of submicron particles may be the result
of inaccurate size distribution determinations of charged
particles with cascade impactors.
89
-------
50
w
H
Q
U
I
w
0.
1Q 20 30 50 70 8
PERCENT LESS THAN STATED SIZE,
0 90 95
Figure 40. Size distribution of particles exitinp
ESP when d =10 ymA, a = 4.0
PgIN g
90
-------
U
z
w
I — I
u
E-
U
,-J
O
u
99.99
99. 9
99.8
99
98
95
90
80
70
60
— THEORETICAL, COMPUTED AT 20 na/cm2
S = 0, a = 0.25
D dpg IN = 10 ymA, a = 0.25
O SRI DATA
0.1
1.0
PARTICLE DIAMETER, ym
8.0
Figure 41. Fractional collection efficiencies for a full-scale
precipitator on a coal-fired power boiler.
(Gooch et al. (1975))
-------
REFERENCES
Calvert, S., Chapter 46 in "Air Pollution", Stern, A.C. (ed).
Third Edition, Academic Press, NY, 1968.
Calvert, S., J. Goldshmid, D. Leith, and D. Metha "Scrubber
Handbook" NTIS PB 213-016, 1972.
Calvert, S., J. Goldshmid, D. Leith, and N. Jhaveri, "Feasibility
of Flux Force/Condensation Scrubbing for Fine Particulate
Collection NTIS PB 227-307, October 1973.
Calvert, S. "Engineering Design of Fine Particle Scrubbers",
J. of APCA. 24-: 929, 1974.
Calvert, S., J. Rowan, and C. Lake "Johns-Manville CHEAP Evalua-
tion" NTIS PB 256-311, July 1975.
Calvert, S., J. Rowan, S. Yung, C. Lake, and H. Barbarika "APS
Electrostatic Scrubber Evaluation" EPA-600/2-76-154a, 1976a.
Calvert, S., C. Christensen, and C. Lake "APS Electro-Tube
Evaluation" PB 258-824, 1976b.
Calvert, S., C. Lake, and R. Parker "Cascade Impactor Calibra-
tion Guidelines" EPA-600/2-76-118, 1976c.
Calvert, S. and R.G. Patterson "Submicron Particle Size Measure-
ment Particle Size Measurement with a Screen Diffusion Battery",
Final Report, EPRI Contract RP 723-1-760205, 1977.
Fegley, M.J., D.S. Ensor, and L.E. Sparks "The Propagation of
Errors in Particle Size Distribution Measurements Performed
Using Cascade Impactors" Paper 75-32.5 presented at the 68th
Annual Meeting of APCA, Boston, MA, June 15-20, 1975.
Harris, D.B. "Procedures for Cascade Impactor Calibration and
Operation in Process Streams" EPA 600/2-77-004, January 1977.
Gooch, J.P. and N.L. Francis "A Theoretically Based Mathematical
Model for Calculation of Electrostatic Precipitator Performance ,
J. of APCA. 2J>: 108-113, 1975.
I par f W W F Krieve, and E. Cohen, "Charged Droplet Scrubbing
for Fine Particle Control," J. of APCA. 25: 184-189, 1975.
92
-------
REFERENCES (continued)
Lipson, C. and N.J. Sheth "Statistical Design and Analysis of
Engineering Experiments", McGraw-Hill, 1973.
Lundgren, D.A. "An Aerosol Sampler for Determination of Particle
Concentration as a Function of Size and Time", J. of APCA 17-
225, 1967. ' —'
Oglesby, S. and G. Nichols "A Manual of Electrostatic Precipi-
tator Technology, Part I. Fundamentals" NTIS BP 196-380, 1970.
Patterson, R.G., P. Riersgard, and S. Calvert "The Effects of
Charged Particles on Cascade Impactor Calibrations", paper
presented at the 70th. Annual Meeting of the APCA, Toronto,
j une Ly 7 7.
Rao, A.K. "Sampling and Analysis of Atmospheric Aerosols", Particle
Technology Laboratory, Mechanical Technology Laboratory, University
of Minnesota. Publication 269, 1975.
S?1^ 5JiJ: ?nd D'W' Co°Per "Laboratory Evaluation of the Cleanable
High Efficiency Air Filter (CHEAP)", EPA 600/2-76-202, 1976.
Sinclair, D. "A Portable Diffusion Battery" American Ind. Hygiene
Assoc. Journal 33: 729-735, 1972.
Smith, W.B., K.W. Gushing, and J.D. McCain "Particle Sizing
Techniques for Control Device Evaluations" EPA 650/2-74-102.
NTIS PB 240670/AS, October 1974.
Sparks, L.E. Personal Communication, 1971.
93
-------
APPENDIX A
CASCADE IMPACTOR DATA AND SIZE DISTRIBUTION
PLOTS FOR TRW CHARGED DROPLET SCRUBBER
94
-------
Table A-l.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN fl
Table A-3.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN 16
10
cn
IMPACTOR
STAGE
NUMBER
Precutter
$ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNmJ)
232
232
232
232
230
180
99.8
50.8
42.9
dpc
(umA)
27.3
12.0
5.7
2.3
1.3
0.74
0.45
0.134
OUTLET
"cum
(mg/DNm3)
35.8
35.8
35.0
34.7
34.0
33.2
27.0
20.2
15.9
dpc
(wmA)
23.8
10.4
4.0
2.0
1.2
0.64
0.37
0.585
Table A-2.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN »4
IMPACTOR
STAGE
NUMBER
Precutter
& Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm')
INLET
M
cum
(mg/DNm3)
163
159
159
159
158
157
120
65.8
28.9
A
V
(ymA)
27.4
12.0
4.6
2.3
1.3
0.74
0.45
---
0.087
OUTLET
M
cum
(mg/DNm3)
61.0
61.0
60.8
60.6
59.8
58.9
45.4
22'. 9
17.2
d
pc
(pmA)
...
23.6
10.3
4.0
2.0
1.2
0.64
0.37
0.475
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
188
187
186
186
186
177
132
67.0
26.4
A
pc
(ymA)
25.6
11.2
5.3
2.2
1.3
0.69
0.42
...
0.099
OUTLET
cum
(mg/DNm3)
11.0
10.8
10.6
10.6
10.6
10.0
7.2
2.8
1.0
DC
(ymA)
— — —
24.3
10.6
4.1
2.1
1.2
0.66
0.38
...
0.602
Table A-4.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN »8
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
247
247
246
245
245
238
194
134
59.0
V
(ymA)
21.6
9.5
4.5
1.8
1.1
0.58
0.3S
.._
0.134
OUTLET
Mcum
(mg/DNm3)
29.2
27.5
26.0
25.3
25.3
25.3
24.9
21.0
14.2
dpc
(ymA)
26.0
11.4
4.4
2.2
1.3
0.70
0.41
---
0.466
-------
ON
Table A-S. INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN 19
Table A-7. INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN *13
IMPACTOR
STAGE
NUMBER
Precutter
3 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm1)
INLET
M
cum
(mg/DNm3)
132
119
113
113
113
113
113
95.4
65.21
V
lumA)
...
24.5
10.7
4.2
2.1
1.2
0.66
0.40
...
0.106
OUTLET
Mcum
(mg/DNm3)
41.5
41.3
40.6
40.6
40.6
40.6
39.5
23.0
13.9
d
(ymA)
26.6
11.6
4.5
2.3
1.3
0.72
0.41
---
0.438
IMPACTOR
STAGE
NUMBER
Precutter
a Nozzle
1
2
3
4
5
6
7
Filter
Sample
Vo lume
(DNm3)
INLET PRIME
M,.
cum
(mg/DNm3)
1789
1678
1651
1636
1569
1247
477
108
27.7
d
pc
(vimA)
...
26.9
11.8
4.6
2.3
1.3
0.73
0.44
...
0.115
INLET
M
cum
(mg/DNm3)
953
945
938
935
923
857
644
233
113
d
PC
(umA)
...
26.6
11.7
4.5
2.3
1.3
0.72
0.42
...
0.088
OUTLET
M
cum
(mg/DNm1)
56.6
54. 8
53.7
52.0
50.2
44.1
29.0
12.5
2.5
d
PC
fumA)
...
23.2
10.2
3.9
2.0
1.1
0.63
0.36
0.279
Table A-6.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN *10
Table A-8.
IMPACTOR
STAGE
NUMBER
Precutter
$ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm1)
INLET PRIME
M
cum
(mg/DNm3)
244
244
244
244
244
244
224
137
57.6
d
pc
(UmA)
...
32.1
14.0
5.4
2.7
1.6
0.87
0.52
0.045
INLET
M
cum
(mg/DNm3)
166
107
105
103
100
100
54.7
29.6
27.4
d
PC
(pmA)
...
24.7
10.8
5.1
2.1
1.2
0.67
0.40
0.044
OUTLET
M
ciun
(mg/DNm3)
84.5
84.5
83.6
82.8
81.6
76.5
48.8
22.4
16.2
,1
PC
(um A)
---
26.0
11.4
4.4
2.2
1.3
0.70
0.40
0.353
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #14
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
201
168
156
146
138
126
90.8
41.4
5.9
V
(ymA)
11.1
4.6
2.6
1.4
0.88
0.48
0.28
...
0.051
OUTLET
M
cum
(mg/DNm3)
23.1
23.1
22.1
21.0
19.5
16.2
10.5
4.4
1.0
dpc
(ymA)
17.3
7.6
2.9
1.5
0.85
0.4-7
0.27
...
0.771
-------
Table A-9.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #15
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
419
410
399
382
360
195
84.5
36.8
7.4
d
PC
(ymA)
---
8.4
4.2
2.4
1.3
0.81
0.45
0.26
...
0.054
OUTLET
M
cum
(mg/DNm3)
35.5
35.2
33.5
31.4
27.2
21.6
12.3
4.9
1.8
d
PC
(ymA)
...
18.8
8.2
3.2
1.6
0.92
0.51
0.29
0.554
Table A-10. INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #16
IMPACTOR
STAGE
NUMBER
Precutter
5 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
182
177
1.74
168
163
149
108
58.0
5.5
A
PC
(pmA)
10.1
4.2
2.4
1.3
0.80
0.44
0.25
0.036
OUTLET
M
cum
(mg/DNm3)
22.2
22.2
21.8
21.2
20.0
17.6
12.9
7.4
1.4
d
pc
(ymA)
_ . .
18.0
7.9
3.1
1.5
0.89
0.49
0.28
0.933
Table A-ll.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #17
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
cum
(mg/DNm3)
508
476
467
456
433
214
119
50.2
9.3
d
pc
(pmA)
9.4
4.8
2.7
1.5
0.90
0.50
0.29
...
0.054
OUTLET
M
cum
(mg/DNm3)
23.1
23.0
22.2
21.1
19.1
15.9
10.3
4.1
1.4
Q
(ymA)
18.6
8.1
3.2
1.6
0.91
0.50
0.29
0.876
Table A-12.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #18
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
168
129
126
121
116
111
103
78.4
24.5
dpc
(ymA)
11.3
4.7
2.7
1.5
0.89
0.49
0.28
...
0.061
OUTLET
M
cum
(mg/DNm3)
21.8
21.8
21.0
19.8
17.5
14.3
9.8
5.2
1.1
dpc
(pmA)
19.2
8.4
3.2
1.6
0.94
0.52
0.30
---
0.943
-------
Table A-13.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #19
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
217
170
166
160
151
135
77.5
50.3
15.0
d
pc
(ymA)
8.7
4.4
2.5
1.4
0.83
0.46
0.27
0.074
OUTLET
M
cum
(mg/DNm3)
25.5
25.5
24.6
23.7
21.9
19.0
13.3
6.3
2.1
dpc
CumA)
19.5
8.5
3.3
1.7
0.96
0.53
0.30
0.663
98
-------
0.2
2 5 10 20 30 40 SO 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, ?
Figure A-l. Inlet and outlet size distribution for run 1.
0.2
2 5 10 20 30 40 SO 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE, %
Figure A-2. Inlet and outlet size distribution for run 4.
-------
5.0
O
o
0.2
2 5 10 20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, I
Figure A-3. Inlet and outlet size distribution for run 6.
4.0
0.2
5 10 20 30 40 SO 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, %
Figure A-4. Inlet and outlet size distribution for run 8.
-------
2 5 10 20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, *
Figure A-5. Inlet and outlet size distribution for run 9.
:
OINLET
INLET PRIME
A OUTLET
0.2
90 95 98
20 30 40 40 60 70 80
MASS PERCENT UNDERSIZE , t
Figure A-6. Inlet and outlet size distribution for run 10.
-------
o
NJ
OINLET
QlNLET PRIME
AOUTLET
0.2
2 5 10 20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, %
Figure A-7. Inlet and outlet size distribution for run 13.
0.2
2 5 10 20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, t
Figure A-8. Inlet and outlet size distribution for run 14.
-------
0.2
10
20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE , 4
Figure A-9. Inlet and outlet size distribution for run 15.
4.0
2 5 10 20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, $
Figure A-10. Inlet and outlet size distribution for run 16.
-------
• [A OUTLET I
'
0.2
2 5 10 20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE, \
Inlet and outlet size distribution for run 17.
Figure A-
2 5 10 20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE, I
Figure A-12. Inlet and outlet size distribution for run 18.
-------
0.1
o
en
0.2
2 5 10 20 30 40 50 60 70 80 90 95 98
MASS PERCENT UNDERSIZE , %
Figure A-13. Inlet and outlet size distribution for run 19.
0.002
5 10 20 30 40 50 60 70 80 90 95 98
NUMBER PERCENT UNDERSIZE, %
Figure A-14. Diffusion battery data for inlet run 1 f, 4.
-------
4 TITLE AMD SUBTITLE
Evaluation of Four Novel Fine Particulate
Collection Devices
1 REPORT NO.
EPA-600/2-78-062
TECHNICAL REPORT DATA
(Please read Inxntctions on the reverse before completing)
5 REPORT DATE
March 1978
[7.AUTHORis)giCalvert, S.C.Yung, H.Barbarika, and
R.G.Patterson
8. PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Air Pollution Technology, Inc.
4901 Morena Boulevard, Suite 402
San Diego, California 92117
12. SPONSORING AGENCY NAME AND ADDRESS
EPA Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
. RECIPIENT'S ACCESSION NO.
i. PERFORMING ORGANIZATION CODE
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-004
11. CONTRACT/GRANT NO.
68-02-1496
14. SPONSORING AGENCY CODE
EPA/600/13
A \,\* fc_< \* t~**> «^ *» •—• — —— C7 ' __ ^—.^^—^^-^^^^^—
is.SUPPLEMENTARY NOTES ffiRL-RTP project officer is Dale L. Harmon, Mail Drop 61,
919/541-2925.
16. ABSTRACTThe report gives results of an experimental performance evaluation of
four novel fine particulate control devices: the Johns-Manville Cleanable High-
EfnC?ency Mr Filtration (CHEAF) System, the APS Electrostatic Scrubber, the
APS Electrotube, and the TRW Charged Droplet Scrubber The evaluations
included measurement of inlet and outlet particle size distribution and concentration
with cascade impactor and diffusion battery. Fine particle collection efficiencies
were computed from the data, as functions of particle size. Mathematical perfor-
Tance models were developed for the CHEAF system and the APS scrubber. The
models gave satisfactory predictions.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air Pollution
Dust Collectors
Evaluation
Measurement
Particle Size Distribution
Mathematical Models
13. DISTRIBUTION STATEMENT
Unlimited
b.lDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
Stationary Sources
Particulate
19. SECURITY CLASS (This Report!
Unclassified .
20 SECURITY CLASS (This pasc)
Unclassified
COSATI Field/Group
13B
13A
14B
12A
21. NO. OF PAGES
120
22. PRICE
v>, i orm 2220-1 (9-73)
106
------- |