EPA-600/2-77-209b
November 1977
Environmental Protection Technology Series
AMERICAN AIR FILTER
KINPACTOR 10x56VENTURI
SCRUBBER EVALUATION
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 five series. These five broad categories were established to
facilitate further development and application of environmental technology. Elimination of
traditional grouping was consciously planned to foster technology transfer and a maximum
interface in related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumenta-
tion, equipment, and methodology to repair or prevent environmental degradation from point
and non-point sources of pollution. This work provides the new or improved technology
required for the control and treatment of pollution sources to meet environmental quality
standards.
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 Information Service,
Springfield, Virginia 22161.
-------
EPA-600/2-77-2Q9b
November 1977
AMERICAN AIR FILTER
KINPACTOR 10 x 56 VENTURI
SCRUBBER EVALUATION
by
Seymour Calvert, Harry Barbarika, and Gary M. Monahan
Air Pollution Technology, Inc.
4901 Morena Boulevard, Suite 402
San Diego, California 92117
Contract No. 68-02-1869
ROAP No. 21ADM-029
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
An American Air Filter Kinpactor 10 x 56 venturi scrubber
operating on the emissions from a large borax fusing furnace has
been evaluated. The average total efficiency was 97.5 I during
the test period. The venturi was operated at a pressure drop of
110 cm W.C., using about 33 liters/s of scrubbing liquor for a
gas flow rate of approximately 20 Am3/s (43,000 CFM) at 80°C. The
dust had a mass median aerodynamic diameter of about 0.8 ymA.
The collection efficiencies of particles with aerodynamic
diameters between 0.3 ymA and 3 ymA were determined from size
distribution data taken with cascade impactors. The efficiency
data showed the venturi to be more efficient than predicted for
particle sizes below 1 ymA. Particle mass augmentation by con-
densed water is a probable reason for the high efficiency for
small particle collection. Diffusion battery data indicate the
occurrence of some particle growth.
Cost data supplied by the user showed that the venturi
scrubber system initially cnst about $29,000/(m3/s) ($13.70/CFM)
in 1970.
This report was submitted in partial fulfillment of Contract
No. 68-02-1869 by Air Pollution Technology, Inc. under the spon-
sorship of the U.S. Environmental Protection Agency.
111
-------
CONTENTS
Abstract iii
Figures v
Tables vii
1. Introduction 1
2. Summary and Conclusions 2
3. Source and Control System 3
4. Test Method 7
5. Process Conditions 10
6. Cascade Impactor Particle Data 13
7. Diffusion Battery Data 17
8. Particle Penetrations 19
9. Opacity 24
10. Economics 25
11. Operating Problems 26
12. Performance Comparison 27
Appendices
A. Size Distribution Data 32
B. Venturi Scrubber Performance Model 43
IV
-------
FIGURES
Number Page
1 Schematic diagram of scrubbing system 4
2 Schematic top view of A.A.F. Kinpactor 10 x 56,
shown in fully open position (dimensions in milli-
meters') 5
3 Schematic of cyclone entrainment separator (dimen-
sions in meters) 5
4 Modified EPA sampling train with in-stack cascade
impactor 3
5 Particle penetration for run 2 20
6 Particle penetration for runs 4 and 5 20
7 Particle penetration for runs 6, 7 and 8 21
8 Particle penetration for runs 9, 10 and 11 21
9 Particle penetrations for runs 12 and 13 22
10 Pa^ ticle penetration for average of runs 12 and 13
compared to prediction \ 30
A-l Inlet and outlet size distribution for run #1 .... 36
A-2 Inlet and outlet size distribution for run #2 .... 35
A-3 Inlet and outlet size distribution for run #4 .... 37
A-4 Inlet and outlet size distribution for run #5 .... 37
A-5 Inlet and outlet size distribution for run #6 .... 38
A-6 Inlet and outlet size distribution for run #7 .... 33
A-7 Inlet and outlet size distribution for run #8 .... 39
A-8 Inlet and outlet size distribution for run #9 .... 39
-------
FIGURES (continued)
Number
A-9 Inlet and outlet size distribution for run #10. . , • 40
A-10 Inlet and outlet size distribution for run #11. . • • 40
A-ll Inlet and outlet size distribution for run #12. . . .
A-12 Inlet and outlet size distribution for run #13. . . . 41
A-13 Size distributions from diffusion battery data. . . . 42
VI
-------
TABLES
Number Page
1 Inlet process conditions 11
2 Outlet process conditions 11
3 Outlet average gas composition 12
4 Particle size distribution summary 16
5 Mass loading and overall penetration 23
6 Opacity 24
A-l Inlet and outlet sample particle data for run #1 . . 33
A-2 Inlet and outlet sample particle data for run #2 . . 33
A-3 Inlet and outlet sample particle data for run #4 . . 33
A-4 Inlet and outlet sample particle data for run #5 . . 33
A-5 Inlet and outlet sample particle data for run #6 . . 34
A-6 Inlet and outlet sample particle data for run #7 . . 34
A-7 Inlet and outlet sample particle data for run #8 . . 34
A-8 Inlet and outlet sample particle data for run #9 . . 34
A-9 Inlet and outlet sample particle data for run #10 . . 35
A-10 Inlet and outlet sample particle data for run #11. . 35
A-ll Inlet and outlet sample particle data for run #12. . 35
A-12 Inlet and outlet sample particle data for run #13. . 35
VII
-------
SECTION 1
INTRODUCTION
Air Pollution Technology, Inc. (A.P.T.) conducted a perfor-
mance evaluation of an American Air Filter Kinpactor 10 x 56 ven-
turi scrubber in accordance with E.P.A. Contract No. 68-02-1869,
"Fine Particle Scrubber Evaluations."
The objective of the performance test was to determine fine
particle collection efficiency as a function of particle size and
scrubber parameters.
Simultaneous inlet and outlet particle sampling measurements
were taken on the scrubber during a six day test period during
August, 1976. Cascade impactors, condensation nuclei counters,
and a portable diffusion battery were used to obtain total mass
loadings and size distribution data. The data and results of the
evaluation of the scrubber are presented in the text.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
The American Air Filter Kinpactor 10 x 56 venturi scrubber
operating on the emissions from a large borax fusing furnace had
an average total efficiency of 97.5 % during the testing period.
The venturi was operated at a pressure drop of 110 cm W.C., using
about 33 liters/s of scrubbing liquor for a gas flow rate of
approximately 20 Am3/s (43,000 ACFM) at 80°C. The dust had a
mass median aerodynamic diameter of about 0.8 ymA.
The collection efficiencies of particles with aerodynamic
diameters between 0.3 ymA and 3 ymA were determined from size
distribution data taken with cascade impactors. The efficiency
data showed the venturi to be more efficient than predicted for
particle sizes below 1 ymA. Probable reasons for the higher ex-
perimental efficiency in the smaller size range of particles were
collection by condensation and particle growth. Diffusion battery
data support the reasoning for the higher experimental efficiency
for the smaller size range. The major condensation mechanism is
that of water vapor on drops which have a low vapor pressure be-
cause they contain dissolved borax (Na2Bi,07).
Cost data supplied by the user showed that the venturi scrub-
ber system initially cost about $29,000/(m3/s) in 1970.
-------
SECTION 3
SOURCE AND CONTROL SYSTEM
The emission source was a large borax fusing furnace used in
continuous operation. The furnace was capable of producing 2.7 x
10s kg per day of anhydrous borax (NaalUOr) from the pentahydrated
form of the feed, but was not always operating at full capacity
during the testing. The particulates emitted were primarily the
hydrated and anhydrous forms of borax which escaped during the
drying and fusing processes.
The total scrubbing system is shown in Figure 1, with the lo-
cation of the sampling ports indicated. The gases from the fur-
nace at a temperature of 1,000°C to 1,100°C are quenched by scrub-
bing liquor to about 80°C. The gases then enter the American Air
Filter Kinpactor 10 x 56 venturi scrubber, which is shown in sche-
matic in Figure 2. The venturi is rectangular in cross section
with a throat height of 142 cm (56 inches). The throat width is
automatically controlled to maintain a pressure drop of about
110 cm W.C. (43 in W.C.) across the venturi. The throat width
can vary from 0.64 cm (0.25 in) to 25.4 cm (10 in). Following
the venturi the gases enter a cyclone entrainment separator which
is shown in Figure 3. The gas is moved through the scrubbing
system by a blower which is rated at 44 Am3/s at 84°C and 114
cm W.C. pressure differential (93,500 ACFM, 184°F, 45 in W.C.).
The blower is powered by a 746 kW (1,000 HP) motor. The gases
exhaust to the atmosphere through a 21 m tall, 2.13 m diameter
stack.
The scrubbing liquor is recycled through a tank which is
fed about 28 H/s of fresh liquor or lesser amounts of fresh wa-
ter. About 22 £/s of concentrated liquor is pumped from the bot-
tom of the tank so that the borax concentration of the liquor is
maintained at from 10 to 15 percent.
-------
To
Atmosphe
From
i ,nnn°r j
T • .1 1 _ , Sampling
Liquor -~*"| j Quencher Port
f1 */f Cyclone
i—i
\
Sampling }
Port Venturi j
20 m3/s V ^Z^\~"" 1
1 ~f 1 1 | f ^V
I 1 J y c j [
lo Liquor To r " I 111™
Recycle 41 l/s Recycle
Tank rmax.) Tank Blower
rel
^Ba^BMMMHH
Stack
70°C
25 m?/s
Figure 1. Schematic diagram of scrubbing system.
-------
Flow Direction
510
254 (max)
710
1830
Figure 2. Schematic top view of A.A.F. Kinpactor 10 x 56, shown in
fully open position (dimensions in millimeters).
-------
2.13
Figure 3. Schematic of cyclone entrainment
separator (dimensions in meters)
-------
SECTION 4
TEST METHOD
The performance characteristics of the American Air Filter
Kinpactor 10 x 56 venturi scrubber were determined by measuring
the particle size distribution and mass loading of the inlet and
outlet gas simultaneously.
For the tests performed in August 1976, modified E.P.A. type
sampling trains with in-stack heated University of Washington
Mark III (U.W.) cascade impactors were used for particle measure-
ments above 0.3 ymA. Figure 4 shows a schematic diagram of the
modified sample train. Glass fiber filter (Gelman type AE) sub-
strates were used in the impactors to prevent particle bounce and ,
minimize wall losses. Low velocity impactor jet stages were used
for the majority of test runs on the inlet sampling to increase
the sampling time.
The Air Pollution Technology portable screen diffusion bat-
tery (A.P .T. -S.D.B .) was used for particle measurements from 0.1
ym to 0.01 ym (actual).
The A.P.T.-S.D.B. uses Brownian diffusion to accomplish the
size fractionation of particles smaller than 0.1 ym. Because
smaller particles diffuse more readily than larger ones, succes-
sively larger particles are captured as they proceed through the
battery.
A condensation .nuclei counter (C.N.C.) was used to determine
the particle number concentration at several locations in the
battery. From this data, the size distribution may be determined
for the particles smaller than 0.1 ym. The size distribution
computation was based on a calibration of the S.D.B. performed
in the A.P.T. laboratory.
During an impactor run, inlet or outlet fine particle size
measurements were taken with the portable diffusion battery.
-------
oo
PRECUTTER
AND
NOZZLE
1
THERMOMETER mpINGER TRAIN
HEATED
CASCADE
IMPACTOR
ORIFICE
METER
7
'l
I
r
L
STACK
'
WALL
'h
fl
•2.
{
MANOMETER
|
THERMOMETERS
_T|ROTAMETER
VACUUM
GAUGE
DRY GAS
METER
VACUUM
PUMP
SILICA
GEL
DRYER
Figure 4. Modified EPA sampling train with in-stack cascade impactor
-------
Since the system remained fairly constant during the test period,
inlet and outlet S.D.B. measurements at different times were con-
sidered to approximate simultaneous sampling.
Impactor blank runs on the inlet were performed periodically
to insure that the 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.
Gas flow rates for all tests were determined by means of a
calibrated standard-type pitot tube along with in-stack taps for
continuous wet and dry bulb temperature measurements. Velocity
traverses of the inlet and outlet were performed according to the
E.P.A. standards and average velocity points selected for one-
point sampling. Sampling flow rates were measured with the usual
E.P.A. train instruments so as to obtain isokinetic sampling.
Orsat analysis of the outlet gases was also performed.
The inlet sample ports were located in the best place avail-
able, but in a relatively poor location for sampling. The inlet
duct was square and 1.52 m (5 ft) on a side. The sample ports
were 0.4 duct diameters downstream of a 90° bend and 1 duct dia-
meter from the beginning of the venturi section. Twenty point
velocity traverses through the center of the duct from both the
top and the side were used to determine the flow rate and the lo-
cation for the cascade impactor sampling. Since the flow was not
well developed at the inlet sampling port, the inlet flow rate
values are less reliable than the outlet flow rates.
The outlet sample ports were located in a 2.13 m (7 ft) dia-
meter round stack, five diameters downstream of the inlet from the
fan and about 4 diameters upstream of the stack exit. The velo-
city traverses indicated well developed flow.
-------
SECTION 5
PROCESS CONDITIONS
Thirteen cascade impactor sampling runs were made over a
period of six days. The gas conditions at the inlet and outlet
sampling locations during the runs are shown in Tables 1 and 2.
It is thought, from observations of the amount and consistency
of the product from the fusing furnace during the period, that
the conditions during runs 9 through 13 are most representative
of normal conditions. The average barometric pressure during
the testing period was 93.63 kN/m2 (27.65 in Hg) .
Close examination of the inlet and outlet flow rates un-
covers a violation of the mass conservation law since the outlet
flow rate is greater than the inlet flow rate while the inlet
temperature is greater than the outlet temperature. The static
pressures are practically equal. There are two explanations for
this flow rate discrepancy: 1) As noted previously, the inlet flow
is not well developed and thus the inlet flow rate is probably
not accurate and, 2) air could be leaking into the system because
of the low pressure (-110 cm W.C. gage) within the system between
the venturi and the blower.
The results on the Orsat analysis of the outlet gas are pre-
sented in Table 3.
10
-------
TABLE 1. INLET PROCESS CONDITIONS
Run
1
2
3,4,5
6,7,8
9,10,11
12,13
Temp.
°C
75
108
73
78
79
81
Water Volume
Percent *
36
57
14
23
49
41
Static Press.
cm W.C.
-8.3
-6.5
-6.1
-7.2
-7.5
-8.6
Flow Rate
Am3/s (ACFM)
24 (51,000)
21 (45,000)
16 (35,000)
18 (38,000)
20 (42,000)
20 (43,000)
TABLE 2. OUTLET PROCESS CONDITIONS
Run
1
2
3,4,5
6,7,8
9 , 10, 11
12,13
Temp.
°C
80
73
72
68
71
69
Water Volume
Percent *
31
33
38
28
23
21
Static Press.
cm W.C.
-0.5
-0.4
-0.4
-0.4
-0.4
-0.4
Flow Rate
Am3/s (ACFM)
29 (62,500)
25 (54,000)
25 (54,000)
25 (54,000)
25 (53,000)
25 (53,000)
* Based on wet and dry bulb temperatures
11
-------
TABLE 3. OUTLET AVERAGE GAS COMPOSITION
Gas Component
N2
02
C02
CO
H20
Molecular Wt .
Volume Percent
Wet
60
6
5
0
29
26
Volume Percent
Dry
85
8
7
0
29
12
-------
SECTION 6
CASCADE IMPACTOR PARTICLE DATA
Particle size distribution data were obtained for the Ameri-
can Air Filter Kinpactor venturi scrubber as described in the Test
Method section. Identical single point sampling at the average
velocity location was performed at both the inlet and outlet. The
sampling time of each run depended on the mass loading. The aver-
age sampling times for the inlet and outlet were nine and fifty
minutes respectively.
Because of the number of very large particles in the inlet
gas and the entrained water droplets at the inlet and outlet, pre-
cutters were used. The aerodynamic cut diameters of the inlet
and outlet precutters were approximately 12 ymA and 4.5 ymA re-
spectively. The inlet sampling was approximately isokinetic.
However, because of the large amount of water droplets in the
outlet gas, a special "rain can" had to be used on the outlet
sampling nozzle. Also, the outlet sampling nozzle was oriented
perpendicular to the gas flow. The velocity through the outlet
nozzle was maintained at the velocity of the outlet flow at that
location.
Isokinetic conditions are not crucial for sampling fine
particles. For example, the error caused by sampling 4 ymA par-
ticles at a velocity 501 higher or lower than the gas stream ve-
locity would only be 2 or 3% of the concentration.
The use of single point sampling and perpendicularly oriented
nozzles is usually permissible when measuring fine particle size
and concentration. The fine particles will be distributed well
in the gas stream, except in cases where streams with different
particle concentrations have not had sufficient time to mix. To
illustrate that one-point sampling is sufficient for fine parti-
13
-------
cles, we may note that Stokes stopping distance of a 3 umA par-
ticle with an initial velocity of 15 m/sec (50 ft/sec) is about
0.04 cm (0.016 inches) and for a 1 ymA diameter particle is one-
ninth of that. Since the stopping distance is the maximum that
a particle can be displaced from a gas streamline by going around
a right angle bend, it is apparent that fine particle distribu-
tion in the gas stream will be negligibly affected by flow direc-
tion changes.
To minimize the possibility of condensation in the impactors
and to collect only the dry particles, the impactors were main-
tained at about 15°C above the gas stream temperature by heating
blankets. The precutters were not heated.
The fact that the impactors were heated should be noted when
interpreting the size distribution data. Some unpublished data
taken with heated and unheated cascade impactors in a stream of
wet borax particles have shown differences in the size distribu-
tions. The primary difference was that the mass median aerody-
namic diameter of the borax particles collected in heated impac-
tors was as low as 70 percent of the mass median aerodynamic
diameter of the particles collected in unheated impactors. Thus,
the venturi scrubber of the present system, which operates on the
principle of impaction, may be experiencing larger aerodynamic
size particles than the heated cascade impactor data indicate.
Data from run 3 are not included because the outlet cascade
impactor stages had become wet. This happened because sampling
with a parallel orientation of the outlet nozzle was tried for
this run.
Particle concentration, particle size and sampled volumes
for cascade impactor runs are tabulated in the Appendix in
Tables A-l through A-12. Size distributions for the impactor runs
are given in Figures A-l through A-12 located in the Appendix. The
inlet and outlet size distributions both indicate a bimodal na-
ture. The straight dashed lines drawn on Figures A-l through
A-12 represent the region of the size distributions where
14
-------
log-normality may exist. The lines on the outlet distributions
may approximate the lower range mode of a bimodal log-normal
distribution. For want of a better ananlytic description of the
size distributions, the mass mean geometric diameters and geomet-
ric standard deviations for the log-normal parameters as well as
the mass median diameter (from the data points) runs are presen-
ted in Table 4.
In this report, the symbol "d " refers to aerodynamic dia-
meter, which is equal to the particle diameter (d ) in microns
(ym) times the square root of the particle density (p ) in grams
per cubic centimeter (g/cm3) times the square root of the Cunning-
ham slip correction factor (C'J. The symbol "ymA" represents the
units of aerodynamic size.
15
-------
TABLE 4. PARTICLE SIZE DISTRIBUTION SUMMARY
Run
1
2
4
5
6
7
8
9
10
11
12
13
Inlet
pm
ymA
0.80
0.94
0.86
0.92
0.82
0.66
0.86
0.76
0.94
0.75
d
PS
ymA
0.84
0.99
2.3
0.77
1.9
0.85
0.79
0.55
0.83
0.69
1.2
0.72
'g
2.1
3.3
3.7
2.7
3.3
4.7
3.3
2.7
3.3
2.7
3.3
3.0
Outlet
pm
ymA
0.40
1.1
0.32
0.33
0.23
0.25
%.
ymA
0.55
0.78
0.32
0.22
0.41
0.20
0.22
0.13
0.16
0.25
0.16
0.19
a
g
2.0
3.2
2.6
2.9
2.1
2.8
3.4
3.6
3.5
5.2
3.6
3.5
Note:
pm
Pg
g
mass median aerodynamic particle diameter from
data
log-normal geometric mass mean aerodynamic
particle diameter
log-normal geometric standard deviation
16
-------
SECTION 7
DIFFUSION BATTERY DATA
Diffusion battery data were taken during the fifth and sixth
days of testing. The inlet data were taken during cascade impac-
tor runs 9, 10 and 11, while the outlet data were taken during
runs 12 and 13. The average diffusion battery data particle size
distribution for the inlet and outlet are shown in Figure A-13 in
the Appendix.
Because of the large amount of water vapor (40% by volume)
in the gas, the lenses in the condensation nuclei counter of the
diffusion battery would fog when the diffusion battery was opera-
ted in the normal manner. Increased dilution did not solve the
problem because the particle count would then drop below the
threshold of the counter. Also, heating the diffusion battery
to its maximum allowable temperature in an effort to reduce conden-
sation on the lenses did not help.
The modification to the system that allowed data to be taken
was to route the incoming source gas through a glass flask before
entering the diffusion battery. Enough of the water vapor con-
densed in the flask so that the condensation nuclei counter did
not become inoperable. Any condensation of water vapor would
cause collection of submicron particles by diffusiophoresis.
Thus, a fraction of the particles did not reach the diffusion
battery.
Since the diffusion battery data were taken with condensa-
tion occurring within the system, the size distributions shown in
Figure A-13 of the Appendix may not be accurate. However, since
the configurations for taking the outlet data and the inlet data
were the same and the process was constant during the testing
period, the data may indicate something about the relation between
the distribution of submicron particles of the inlet and outlet.
The data do show that the outlet submicron particles are larger
17
-------
and more monodisperse.
The data showing particle growth are only qualitative since
condensation occurred within the measuring system. Mechanisms
for particle growth are present in the scrubber system. These
mechanisms are discussed in the section on performance compari-
sons .
18
-------
SECTION 8
PARTICLE PENETRATIONS
Particle penetration versus particle aerodynamic diameter
was calculated from the cascade impactor data. The results are
shown in Figures 5 through 9 for each day of testing. Penetra-
tions for run 1 are not shown because they are much higher than
the penetrations for all the other runs, indicating anomalous
behavior. Penetrations for runs of the last two days, when opera-
tions were smoothest, are quite consistent.
Because the size distributions were not log-normal all of
the penetrations were calculated manually. The manual method in-
volved visually determining the slope of the cumulative mass load-
ing versus aerodynamic particle diameter curve, drawn from the
data presented in the Appendix. The ratio of the slopes of the
outlet curve to the inlet curve at a certain particle size was
the penetration for that particle size.
Particle penetration based on diffusion battery data is not
presented because of the inaccuracies incurred during the data
acquisition. Diffusion battery data are discussed in the previous
section.
The total mass loadings and overall penetrations for the
runs are presented in Table 5. The total mass loading was deter-
mined from analysis of the cascade impactor data. Run 1 had ano-
malously high overall penetration and high flow rates which indi-
cated that something may have been wrong with the data or that
the venturi was not operating properly during that run. The
average penetration for all of the runs, exclusive of run 1, was
2.51.
As noted previously, the size distributions were obtained
from analysis of heated cascade impactor data. The effect of this
heating was discussed in the section on cascade impactor data.
19
-------
0.1
ts)
o
0.04
§
p-l
o
0.01
ft. 0.004
0.001
0.3 0.5
PARTICLE AERODYNAMIC DIAMETER, dpa, ymA
I
Figure 5. Particle penetration for run 2.
0.001
0.3 0.5
PARTICLE AERODYNAMIC DIAMETER, d a, ymA
pa
Figure 6. Particle penetration for runs
4 and 5.
-------
0.04
2
O
i—i
H
U
o
I-H
H
W
2
o.oi ;;
o.oo
o.i
0.004
0.5
PARTICLE AERODYNAMIC DIAMETER, d , jimA
pa
Figure 7. Particle penetration for runs 6, 7,
and 8.
0.04
O
H
i
PL,
o 0.01
H
W
W
0.004
0.001
0.3 0.5 1 23
PARTICLE AERODYNAMIC DIAMETER, d^-umA
pa
Figure 8. Particle penetration for
runs 9, 10 and 11.
-------
0.001
imilMIIIIIIIIIII IIIMIIIMMIIIIIIIIIIII
0.3 0.5
PARTICLE AERODYNAMIC DIAMETER, d , umA
pa
Figure 9. Particle penetrations for runs 12 and 13.
22
-------
TABLE 5. MASS LOADING AND OVERALL PENETRATION
Run
1
2
4
5
6
7
8
9
10
11
12
13
Inlet Mass Loading
mg/DNm3
332
653
1,150
841
690
2,310
882
829
1,040
863
1,210
1,000
Outlet Mass Loading
mg/DNm3
40.4
18.3
42.1
11.0
21.9
18.4
20.5
22.7
19.3
31.0
36.9
27.3
Penetration
%
12.2
2.8
3.7
1.3
3.2
0.8
2.3
2.7
1.9
3.6
3.0
2.7
23
-------
SECTION 9
OPACITY
Opacity for the outlet stack of the American Air Filter
Kinpactor 10 x 56 venturi scrubber was determined by personnel
trained and certified by the California Air Resources Board.
t
Readings were made hourly during the testing periods. The opa-
city determinations were made somewhat difficult by the presence
of steam condensation in the plume and the proximity of other
stacks emitting similar particulates.
Table 6 presents the daily average opacity readings.
Date
8/19/76
8/20/76
8/21/76
8/22/76
8/23/76
TABLE 6. OPACITY
Runs_ Average Opacity, %
2 10-15
3-5 15
6-8 20
9-11 15-20
12,13 20
Average Outlet
Loading,
9
12
11
14
19
24
-------
SBCTION 10
ECONOMICS
Data for the initial costs of the venturi-cyclone scrubbing
system, purchased in 1970, were supplied by the user:
Approximate scrubber purchase cost $200,000
Scrubber auxiliaries:
1, Fans, motors, etc. 30,000
2. Ducting 47,000
3. Liquid and solid handling and
treatment 50,000
4. Instrumentation 45,000
5. Electrical material 36,000
Scrubber installation:
1. Site preparation 108,000
2. Installation 100,000
3. Start-up and modification 51,000
4. Engineering 63,000
Total Initial Cost - $730,000
The operating costs were not available. However, the power
costs can be estimated. The major power user is the large blower,
The blower is powered by a 746 kW (1,000 HP) motor. At $0.03 per
kW-hour the fan would cost $22 per hour, or $537 per day.
25
-------
SECTION 11
OPERATING PROBLEMS
The primary operating problems with the system are the plug-
ging of nozzles and piping and scale build-up in the system. These
problems are all caused by calcium carbonate and sulfate deposits.
The local water is very hard and the feed to the fusing furnaces
may also contain these mineral impurities. To combat this prob-
lem special reamer nozzles have to be used and the system has to
be shut down periodically to chip away the built-up scale.
Although a large amount of entrained water was carried-over
from the entrainment separator through the fan, and out the stack,
it was not considered a problem by operating personnel.
26
-------
SECTION 12
PERFORMANCE COMPARISON
The performance of venturi-type scrubbers has been modeled
extensively. The most recent survey and model* are presented in
the Appendix and used here.
The scrubber parameters are known or estimated as follows:
Pressure drop 110 cm W.C.
Gas flow rate 20 Am3/s
Liquid flow rate 0.033 m3/s (QL/QG=0.00171
Maximum throat area 0.361 m2
Minimum throat velocity 69 m/s
Gas temperature 75°C
Gas density 0.79 kg/m3
Gas viscosity 1.6 x 10" kg/m-s
Liquid density 1,000 kg/m3
The gas flow rate chosen was based on outlet flow rate for
runs 12 and 13 and the liquid flow rate was taken at 80 percent
of the stated flow rate since it is believed that because of plug-
ging the maximum rate was not maintained.
From the known pressure drop and a value of 0.8 for the re-
covery factor in the pressure equation, the following are derived:
Throat gas velocity 89 m/s
Throat area 0.28 m2
Liquid drop diameter 0.0104 cm
Using the equations presented in Appendix B a predicted
penetration versus particle aerodynamic diameter curve was con-
*Calvert, S., S. Yung and H.F. Barbarika, "Venturi Scrubber Per-
formance Model," A.P.T., Inc., San Diego, CA, EPA Contract #68-02-
1328, Task No. 13, July 1976.
27
-------
structed. This curve is the "prediction" curve on Figure 10.
In order to compare the model with experiment a few important
factors need to be noted:
1. Heated impactors - As explained previously the inlet and
outlet size distributions were determined using heated cascade
impactors. Based on previous experience it was expected that
the mass median diameters of the actual wet borax particle size
distributions were up to 1.5 times the sizes measured in the
heated impactors. The high solubility of borax was the primary
reason for the difference. Thus the venturi was collecting lar-
ger particles than those measured in the heated impactor. In
order to compare the experimental results with the prediction
it is necessary to plot experimental penetration against the
actual (wet) particle size rather than the dried particle size.
This correction causes the measured penetration curve in Figure
10 to shift toward larger particle diameters.
2. System leaks - The actual scrubber penetrations could
also be different from those based on the impactor data because
of dilution of the outlet by air leaks into the system, which
probably did occur. Dilution of the outlet stream would cause
the actual penetrations to be greater than those measured by a
factor equal to the dilution factor. The outlet concentration
may have been diluted as much as 25%. The dashed line on Figure
10 includes the effects of 25% dilution and 1.5 times particle
growth on measured data.
3. Collection mechanism - The model for ventrui performance
assumes particle collection by inertial impaction on drops in
the throat region of the venturi. This model does not account
for forces other than inertia which can effect submicron parti-
cle collection. Thus, in the region below about 0.5 ymA the
predicted penetration and the experimental penetration are expec-
ted to differ.
Another important phenomenon is the collection of particles
of all sizes due to solution induced condensation. Borax (NaaB
is highly soluble in water. If the scrubber liquor became more
concentrated than would be in equilibrium with the vapor, it is
28
-------
conceivable that particles could be swept toward the liquid as
water vapor moves to condense on it. The liquid drops which con-
tain dissolved borax have a reduced vapor pressure which in-
duces the condensation which causes the growth in particle size
manifested in the differences between wet and dry particle
sizes .
Some of this dissolution and condensation will have occurred
in the quench section upstream of the venturi. However, because
of the large amount of liquid injected at the venturi these
mechanisms may still be causing particle collection and growth
downstream of the venturi. Growth or collection of the submicron
particles after the inlet sampling point, either before or after
the venturi would explain some of the differences between the
predicted and the experimental penetrations.
Condensation can also occur when the vapor pressure of the
water is reduced because it is at a lower temperature than the
adiabatic saturation temperature of the gas. This cause of con-
densation was not significant here because there was very little
cooling of the scrubber liquor.
4. Sensitivity of prediction to L/G ratio - The predicted
curve is very sensitive to the liquid to gas volume flow rate
ratio (L/G) for particle diameters above 1.0 ymA. A 25% decrease
in the L/G ratio would cause the prediction to agree closely
with the data. Since neither the inlet gas flow rate nor the
liquid flow rate are known precisely, it is possible that the
actual L/G ratio was lower than the value used in the prediction.
5. Entrainment - The carry-over from the cyclone was so heavy
that precutters had to be used on the outlet sampling probe. It
was possible that some smaller entrainment drops, containing
dissolved borax, penetrated the precutters or were shattered in
the precutters and were collected in the heated impactors. Since
the venturi model assumed that no entrainment carry-over occur-
red, the actual penetrations would be greater than the model
prediction, as seen in Figure 10.
These five factors help explain the differences between the
29
-------
0.10
0.04
H
U
o 0.01
i—i
H
0.004
0.001
Prediction
Corrected for
Wet Particles
With Dilution
Measured for
Dry Particles
0.3 0.5
PARTICLE AERODYNAMIC DIAMETER, d , ymA
pa
Figure 10. Particle penetration for average of runs
12 and 13 compared to prediction.
30
-------
model prediction and the data. The many factors and uncer-
tainties involved are enough to preclude any judgment of the
accuracy of the model.
31
-------
APPENDIX A
SIZE DISTRIBUTION DATA
32
-------
TABLE .VI. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN * 1
IMP ACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
7
Filter
Sample
Volume
CDNm3)
INLET
M
cum
(mg/DNm3)
332
312
273
273
263
228
74.4
39.7
39.7
d
Pc
CumA)
*
35.8
3.06
1.75
0.97
'0.58
0.32
0.19
0.020
OUTLET
M
cum
Cmg/DNm3)
40.4
37.3
36.6
35.9
34.7
33.7
31.8
23.9
4.45
d
Pc
OmA)
*
16.1
7.06
2.73
1.38
0.79
0.44
0.25
0.584
TABLE A-2. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN *2
IMPACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
7
Filter
Sample
Volume
CDNm3)
INLET
Mcum
Cmg/DNm3)
653
458
458
447
422
284
165
85.2
45.4
V
CumA)
30.0
2.57
1.47
0.81
0.49
0.27
0.16
0.053
OUTLET
Mcum
Cmg/DNm3)
IS. 5
14.3
11.6
10.7
10.4
9.56
8.42
6.99
3.00
V
CvmA)
16.7
7.31
2.83
1.43
0.82
0.45
0.26
0.701
The inlet and outlet d 's for the precutters averaged
approximately 12 ymA and 4.5 umA respectively for all runs.
TABLE A-3. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #4
IMPACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
d
7
Filter
Sample
Volume
(DNm3)
INLET
"CUK
(mg/DNm3)
1500
524
51?
498
461
360
218
113
46.1
d
pc
(pmA)
44.3
3.79
2.17
1.20
0.72
0.40
0.23
0.024
OUTLET
Mcum
Cmg/DNm3)
42.1
38.3
32.7
31.9
31.7
30.3
29.6
27.5
18.2
V
CumA)
17.2
7.51
2.91
1.47
0.84
0.46
0.27
0.788
TABLE A-4. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #5
TMPAfTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
7
Filter
Sample
Volume
CDNm3)
INLET
Mcum
Cmg/DNm3)
841 -
714
714
708
692
602
320
181
107
d
pc
CvmA)
41.6
3.56
2.04
1.12
0.68
0.37
0.22
0.019
OUTLET .
Mcum
Cmg/DNm3)
11.0
10.0
10.0
10.0
10.0
10.0
9.63
8.23
5.83
V
CumA)
16.1
7.04
2.72
1.37
0.79
0.43
0.25
0.789
-------
TABLE A-S. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN 06
IMP ACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M_ ,m
cum
(mg/DNm3)
691
417
293
293
283
233
159
34.8
24.8
d
pc
CvmA)
38.8
3.32
1.90
1.05
0-.63
0.35
0.20
0.020
OUTLET
M
cum
Cmg/DNm3)
21.9
20.0
18.9
18.2
17.4
16.7
15.3
13.3
4.58
d
V
CumA)
14.7
6.42
2.48
1.25
0.72
0.40
0.23
0.873
TABLE A-6. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #7
TMPAfTOR
STAGE
NUMBER
Precutter
1
2
3
4
7
Filter
Sample
Volume
(DNm5)
INLET
Mcum
(mg/DNm3)
2310
2040
1910
1820
1720
1420
811
678
506
d
PC
CvmA)
40.8
3.49
2.00
1.10
0.67
0.37
0.21
0.010
OUTLET
Mcum
(mg/DNm3)
18.4
17.9
17.3
17.0
16.3
15.7
14.7
13.8
10.0
V
(pmA)
15.1
6.59
2.55
1.29
0.74
0.41
0.23
0.767
TABLE A-7. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #8
IMPACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
882
821
787
749
710
621
338
227
128
d
pc
CumA)
40.8
3.49
2.00
1.10
0.67
0.37
0.21
0.018
OUTLET
M
cum
(mg/DNm3)
20.5
18.5
18.5
18.5
18.5
18.0
16.9
14.4
10.3
d
PC
CumA)
14.6
6.41
2.48
1.25
0.72
0.40
0.23
0.390
TABLE A-8. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #9
IMPACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm1)
INLET
Mcum
(mg/DNm3)
829
796
784
778
739
596
332
183
129
d
PC
CumA)
32.7
2. -80
1.61
0.89
0.53
0.29
0.17
0.033
OUTLET
Mcum
(mg/DNm3)
22.7
21.7
21.6
21.3
21.1
20.6
20.4
18.8
14.8
V
CvmA)
14.2
6.23
2.41
1.22
0.70
0.38
0.22
0.695
-------
TABLE A-9. INLET AND OUTLET SAMPLL PARTICLE DATA FOR RUN #10
TABLE A-ll.INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN 112
IMP ACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
, 7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
1040
885
885
859
792
549
305
189
116
d
pc
CumA)
32.4
2.77
1.59
0.88
0.53
0.29
0.17
0 . 0 34
OUTLET
M
cum
(mg/DNm3)
19.3
18.7
18.7
18.7
18.7
18.3
17.2
14.8
11.5
a
v
(limA)
14.4
6.32
2.44
1.23
0.71
0.39
0.22
0.948
IMP ACTOR
STAGE
NUMBER
Precutter
I
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
Mcum
(mg/DNm3)
1210
912
898
875
812
647
296
155
91.2
V
(vmA)
35.8
3.06
1.76
0.97
0.58
0.32
0.19
0.022
OUTLET
Mcum
(mg/DNm3)
36.9
36.0
35.5
33.8
32.8
32.0
30.6
28.0
22.4
V
CumA)
14.9
6.52
2.52
1.27
1.73
0.40
0.23
0.903
1/1
TABLE A-10. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #11
TABLE A-12. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #13
IMPACTOR
STAGE
NUMBER
Precutter
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
Mcum
(mg/DNm3)
863
771
771
766
735
500
273
150
99.1
d
pc
CumA)
31.8
2.72
] .56
0.86
0.52
0.29
0.17
0.041
OUTLET
Mcum
(mg/DNm3)
31.0
27.6
27.0
26.3
25.3
24.2
22.5
19.3
14.5
V
CumA)
14.5
6.35
2.46
1.24
0.71
0.39
0.23
0.928
IMP ACTOR
STAGE
NUMBER
Precutter
1
Z
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
Mcu»
(mg/DNm3)
1000
912
905
883
832
647
371
207
134
d
PC
CumA)
35.8
3.06
1.76
0.97
0.58
0.32
0.19
0.028
OUTLET
Mcum
Cmg/DNm3)
27.3
25.7
25.5
25.2
24.2
23.2
22.1
19.5
15.2
Sc
CumA)
14.5
6.34
2.45
1.24
0.71
0.39
0.23
0.612
-------
5.0
0.1
10
20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90 95
5.0
0.1
10
20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
95
Figure A-l. Inlet and outlet size distribution for run »1.
Figure A-2. Inlet and outlet size distribution for Run #2.
-------
5.0
0.1
10 20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90 95
Figure A-3. Inlet and outlet size distribution for run #4.
5.0
0.1
20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90 95
Figure A-4. Inlet and outlet size distribution for run 15.
-------
0.1
20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90 95
S.O
1 2.0
w 1.0
S 0.5
o
8
0.2
0.1
10
20 30 40 50 60 70 80
MASS PERCENT UNDBRSIZE
90 95
Figure A-5. Inlet and outlet size distribution for run #6.
Figure A-6, Inlet and outlet size distribution for run #7.
-------
5.0
0.1'
10 20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90 95
Figure A-7. Inlet and outlet size distribution for run
#8.
0.1
10 20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90 95
Figure A-8. Inlet and outlet size distribution for run #9.
-------
5.0
5.0
0.1
10
20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
95
0.1
10 20 30 40 50 60 70 80 90 95
MASS PERCENT UNDERSIZE
Figure A-9. Inlet and outlet size distribution for run #10.
Figure A-10. tnlfet and outlet size distribution for run #11.
-------
s:o
2.0
nl
a.
OS If)
ta 1-u
H
§
0.2
0.1
10 20 30 40 50 60 70 80
m.SS PERCENT UNDERSIZE
90 95
Figure A-ll. Inlet and outlet size distribution for run #12.
s.o m
0.1
20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90
95
Figure A-12. Inlet and outlet size distribution for run #13.
-------
0.3
0.02
20 50 80
NUMBER PERCENT UNDERSIZE
95
Figure A-13. Size distributions from diffusion battery data.
42
-------
APPENDIX B
VENTURI SCRUBBER PERFORMANCE MODEL
43
-------
VENTURI SCRUBBER PERFORMANCE MODEL
Calvert, et al., 1976 have performed a literature
review and evaluation of all available venturi scrubber
performance models. Their conclusions and recommended
performance model are presented below.
(1) Even though each investigator presented a dif-
ferent equation for the prediction of particle collection
in a venturi scrubber, most of these equations can be
reduced to the same basic model, i.e.,
fz 3 u Q n
-in Pt (d ) - / 2 u (; I j . dz (B-l)
o G G r d
where Pt(d ) = penetration for particles with diameter d ,
fraction
u = relative velocity between dust and drop,
cm/sec
UG = gas velocity, cm/sec
d^ = drop diameter, cm
n = single drop collection efficiency, fraction
Q - liquid volumetric flow rate, cm3/sec
z = length, cm
(2) A generalized method for applying equation B-l to
predict particle collection in a venturi was developed.
(3) Particle collection predicted by equation B-l
agrees satisfactorily with performance data.
(4) Most of the particle collection occurs in
the venturi throat. The solution to equation B-l for the
venturi throat, using the inertial collection efficiency
correlation, and assuming a zero initial drop velocity, is
44
-------
B
-[•
- 5.02
po \ d. i.
po
t 1 + f\ 7 4 K™ + 4-2 - 5.02
Kpo + *177 L P°
(B-2)
where u - 2 l -
Pt(d ) = penetration for particles with diameter d ,
fraction
B =
Q. = volumetric liquid flow rate, cm3/sec
PT = liquid density, g/cm3
PG = gas density, g/cm3
Cn = drag coefficient obtained from the "standard curve"
Do °
u5 - dimensionless drop velocity
u,
= »Gt
u, = drop velocity, cm/sec
UG = gas velocity in the throat, cm/sec
K = inertial parameter based on throat velocity
- C' dP PP "Gt
9 yr d,
b u
45
-------
C1 = Cunningham slip factor
d = particle diameter, ym
p - particle density, g/cm3
PP = gas viscosity, poise
d, = drop diameter, cm
L = dimensionless throat length
L = 3 *t So PG
2 dd PL
Equation B-2 slightly under estimates the particle
collection occurring in a venturi scrubber. For most
industrial venturi scrubbers, particle collection can be
predicted closely by neglecting the first term in the
right hand side of equation B-2.
(5) Pressure drop predictions by the modified Calvert's
equation and by Boll's equation agree with experimental data.
The modified Calvert's equation has the following form,
AP = 1.03 x 10 3 Fj u * «i (B-3)
where AP = pressure, cm W.C.
uGt = gas velocity i-n tne throat, cm/sec
Qj = liquid flow rate, cm3/sec
QG = gas -flow rate, cm3/sec
F! = correction factor, dimensionless
Fl = = 2 l - X2+ &- X)
20'5
uGt
v * *tCDo PG . ,
16 d P
46
-------
ude = drop velocity at the exit of the throat, cm/sec
5, = throat length or distance between liquid
injection point and the exit of throat, cm
d, = drop diameter, cm
PG = gas density, g/cm3
PL = liquid density, g/cm3
C~ = drag coefficient at the liquid injection point.
(6) The use of a drag coefficient from the "Standard
curve" gives a better fit between model and experimental
data than does Ingebo's correlation.
(7) The drop diameter can be assumed to be the Sauter
mean diameter calculated from the Nukiyama-Tanasawa relation.
Reference: Calvert, S., S. Yung and H.F. Barbarika, "Venturi
Scrubber Performance Model," A.P.T., Inc., EPA
Contract No. 68-02-1328, Task 13, July 1976.
47
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-5::/2-77-2C9b
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
American Air Filter Kinpactor 10 x 56 Venturi
Scrubber Evaluation
5. REPORT DATE
November 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Seymour Calvert, Harry Barbarika, and
Gary M. Monahan ^
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Air Pollution Technology, Inc.
4901 Morena Boulevard, Suite 402
San Diego, California 92117
10. PROGRAM ELEMENT NO.
IAB012; ROAP 21ADM-029
11. CONTRACT/GRANT NO.
68-02-1869
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
13. TYPE OF REPORT AND
Final; 8/76-10/77
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES IERL-RTP project officer for this report is Dale L. Harmon, Mail
Drop 81, 919/541-2925.
16. ABSTRACT
The report gives results of an evaluation of an American Air Filter Kinpac-
tor 10 x 56 venturi scrubber, operating on emissions from a large borax fusing fur-
nace. Average total efficiency was 97. 5% during the test period. The venturi was
operated at a pressure drop of 110 cm W.C. , using about 33 liters/s o.f scrubbing
liquor for a gas flow rate of about 20 A cu m/s (43,000 CFM) at 80 C. The dust
had a mass median aerodynamic diameter of about 0.8 micrometers A. The collec-
tion efficiencies of particles with aerodynamic diameters between 0.3 and 3 micro-
meters A were determined from size distribution data taken with cascade impactors.
The efficiency data showed the venturi to be more efficient than predicted for par-
ticle sizes below 1 micrometer A. Particle mass augmentation by condensed water
is a probable reason for the high efficiency for small particle collection. Diffusion
battery data indicate the occurrence of some particle growth.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COS AT I Field/Group
Air Pollution
Scrubbers
Venturi Tubes
Borax
Fusion (Melting)
Furnaces
Dust
Air Pollution Control
Stationary Sources
Venturi Scrubbers
Kinpactor
Fusing Furnace
13B
07A
14B
08G,07B
20M
13A
11G
3. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
56
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 22ZO-1 (9-73)
48
------- |