EPA-600/2-77-209a
October 1977
Environmental Protection Technology Series
GAS-ATOMIZED SPRAY
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 Protec-
tion Agency, have been grouped into nine series. These nine broad categories were
established to facilitate further development and application of environmental tech-
nology. 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 TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumen-
tation, equipment, and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the new or improved tech-
nology required for the control and treatment of pollution sources to meet environmental
quality standards.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved for
publication. Approval does not signify that the contents necessarily reflect the views and
policies of the Government, 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-600/2-77-209a
October 1977
GAS-ATOMIZED SPRAY
SCRUBBER EVALUATION
by
Seymour Calvert, Harry F. 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
Fine particle collection efficiency measurements were made
on a gas atomized spray scrubber which cleaned the effluent gas
from a No. 7 gray iron cupola. Tests were made at several levels
of pressure drop and liquid/gas ratio. Particle size measure-
ments on inlet and outlet gas streams were made with cascade im-
pactors and an A.P.T. screen type diffusion battery.
The particle mass collection efficiency at a pressure drop
of about 100 cm W.C. was 911 for particles with a mass median
diameter of about 0.4 ymA. Scrubber inlet gas flow rate could
vary because the cupola was operated with an open top and it
ranged from 4.0 to 12.6 Am3/s ( 8,500 to 27,000 ACFM). Air
leakage into the scrubber system caused serious operating prob-
lems until most of the leaks were sealed.
The penetrations for 1 ymA diameter particles were about as
predicted by the mathematical model, while smaller particle pene-
trations were lower and larger particle penetrations higher than
predicted. The latter two effects are believed due to water con-
densation effects and entrainment, respectively.
This report was submitted in partial fulfillment of Contract
No. 68-02-1869 by Air Pollution Technology, Inc. under sponsor-
ship of the U.S. Environmental Protection Agency.
111
-------
TABLE OF CONTENTS
Abstract ii:L
List of Figures v
List of Tables vii
Sections
Introduction 1
Summary and Conclusions 2
Source and Control System 4
Test Method 7
Process Conditions 10
Cascade Impactor Particle Data 15
Diffusion Battery Data 20
Particle Penetrations 22
Operating Problems 32
Economics 34
Performance Comparison 35
References 45
Appendix A - Cascade Impactor Data 46
Appendix B - Weibull Distribution 60
Appendix C - Venturi Performance Model 64
IV
-------
LIST OF FIGURES
No_. Page
1 Schematic Drawing of Scrubber System 5
2 Modified E.P.A. Sampling Train with
In-Stack Cascade Impactor 8
3 Particle Penetration for Cascade Impactor
Runs 1 and 2 25
4 Particle Penetration for Cascade Impactor
Runs 3 and 4 26
5 Particle Penetration for Cascade Impactor
Runs 5, 6 and 7, and Diffusion Battery
Runs 3-9 (inlet) and 10-12 (outlet) 27
6 Particle Penetration for Cascade Impactor
Run 8 28
7 Particle Penetration for Cascade Impactor
Runs 9 and 10 29
8 Particle Penetration for Cascade Impactor
Runs 11, 12 and 13, and Diffusion Battery
Runs 15 and 16 (inlet) and 13 and 14
(outlet) 30
9 Particle Penetration for Cascade Impactor
Runs 14 and 15 31
10 Comparison of Predicted with Measured
Penetration for Average of Runs 5, 6, and
7 , . 38
11 Comparison of Predicted with Measured
Penetration for Run 8 39
12 Comparison of Predicted with Measured
Penetration for Average of Runs 9 and 10 ... 40
13 Comparison of Predicted with Measured
Penetrations for Average of Runs 11, 12,
and 13 41
14 Comparison of Predicted with Measured
Penetration for Average of Runs 14 and 15. . . 42
A-l Inlet and Outlet Size Distribution
for Run 1 52
A-2 Inlet and Outlet Size Distribution
for Run 2 52
-------
FIGURES (CONTINUED)
No.
A-3
A-4
A-5
A- 6
A- 7
A- 8
A- 9
A-10
A-ll
A-12
A-13
A-14
A-15
Inlet and
for Run 3
Inlet and
for Run 4
Inlet and
for Run 5
Inlet and
for Run 6
Inlet and
for Run 7
Inlet and
for Run 8
Inlet and
for Run 9
Inlet and
for Run 10
Inlet and
for Run 11
Inlet and
for Run 12
Inlet and
for Run 13
Inlet and
for Run 14
Inlet and
for Run 15
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Size
Size
Size
Size
Size
Size
Size
Size
Size
Size
Size
Size
Size
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Distribution
Page
^•^ —
... 53
... 53
... 54
... 54
... 55
... 55
... 56
... 56
... 57
... 57
... 58
... 58
. . . 59
VI
-------
LIST OF TABLES
No. Page
1
2
3
4
5
6
7
8
9
Scrubber Conditions 11
Process Conditions 12
Average Gas Composition 14
Cascade Impactor Data Using Log- Probability
Analysis 18
Cascade Impactor Data Using Weibull Analysis. . 19
Diffusion Battery Particle Size
Distributions 21
Mass Loading and Overall Penetration 24
Average Conditions for Performance
Predictions 36
Predicted Pressure Drop and Overall
Penetration
A-l
A- 2
A- 3
A- 4
A- 5
A- 6
A- 7
A- 8
A-9
A-10
Inlet and
for Run 1
Inlet and
for Run 2
Inlet and
for Run 3
Inlet and
for Run 4
Inlet and
for Run 5
Inlet and
for Run 6
Inlet and
for Run 7
Inlet and
for Run 8
Inlet and
for Run 9
Inlet and
for Run 10
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Outlet
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Particle
Particle
Particle
Particle
Particle
Particle
Particle
Particle
Particle
Particle
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
. . 37
. . 47
. . 48
. . 48
. . 48
. . 48
. . 49
. . 49
. . 49
. . 49
. . 5Q
Vll
-------
TABLES (CONTINUED)
N£- Page
A-ll Inlet and Outlet Sample Particle Data rn
for Run 11 ................... 5U
A-12 Inlet and Outlet Sample Particle Data ,-n
for Run 12 ................... s
A-13 Inlet and Outlet Sample Particle Data
for Run 13 ................... bu
A-14 Inlet and Outlet Sample Particle Data
for Run 14 ................... bi
A-15 Inlet and Outlet Sample Particle Data
for Run 15 ................... 51
Vlll
-------
SECTION 1
INTRODUCTION
Air Pollution Technology, Inc. (A.P.T.) conducted a
performance evaluation of a gas-atomized spray scrubber in
accordance with EPA Contract Number 68-02-1869 "Fine Par-
ticle Scrubber Evaluation". The scrubber controlled emissions
from a No. 7 gray iron cupola.
The objective of this performance test was to determine
fine particle collection efficiency as a function of particle
size and scrubber operating parameters.
Simultaneous inlet and outlet particle sampling measure-
ments were taken on the scrubber during a two-week test period
from February 21 to March 4, 1977. Cascade impactor, conden-
sation nuclei counters, and a portable diffusion battery were
used to obtain total mass loadings and size distribution data.
The data and results of this scrubber evaluation are presented
in this text.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
Fine particle collection efficiency measurements were made
on a gas atomized spray scrubber which cleaned the effluent gas
from a No. 7 gray iron cupola. Tests were made at several levels
of pressure drop and liquid/gas ratio.
The scrubber throat was a rectangular nozzle with a variable
outlet width. Hot gas from the cupola passed through an after-
burner and then a water spray quencher before reaching the scrubber
throat. A cyclone separator located downstream from the scrubber
removed liquid from the gas and returned it to the scrubber sump,
from which it was pumped back to the scrubber throat.
The penetrations for 1 ymA diameter particles were about as
predicted by the mathematical model, while smaller particle pene-
trations were lower and larger particle penetrations higher than
predicted. The latter two effects are believed due to water con-
densation effects and entrainment, respectively.
The particle mass collection efficiency at a pressure drop of
about 100 cm W.C. was 91% for particles with a mass median diameter
of about 0.4 ymA. Scrubber inlet gas flow rate could vary because
the cupola was operated with an open top and it ranged from 4.0 to
12.6 Am3/s (8,500 to 27,000 ACFM). Air leakage into the scrubber
system caused serious operating problems until most of the leaks
were sealed.
Particle size measurements on inlet and outlet gas streams
were made with cascade impactors and an A.P.T. screen type dif-
fusion battery. Liquid entrainment was observed in the outlet
gas and is evidenced by an appreciable amount of mass collection
in the pre-cutter.
The performance tests were done during a period of cold
(around 0°C) and windy weather. This caused condensation to occur
-------
through the mechanisms of heat transfer from the gas and liquid
through the vessel and ducting walls and gas temperature re-
duction due to cold air leakage. Consequently the collection
efficiency for sub-micron particles was enhanced.
Corrosion of steel vessels and ducting was caused by the
absorption of sulfur oxides and collection of salts from the
cupola gas in scrubber liquid. Sodium carbonate addition to
the scrubber liquid was begun during the test period in order to
keep the pH near 7.0 and thereby reduce corrosion. Regularly
scheduled cleaning of the scrubber sump was also started in or-
der to reduce solids carry-over with entrainment.
Following the tests, the cupola was outfitted with a water-
sealed cap and a closure for the charging door. With these modi-
fications the scrubber is capable of satisfactorily controlling
emissions. The entrainment separator efficiency is inadequate
and it should be either improved or replaced.
Computation of penetration as a function of particle size
was attempted with the use of the Weibull relationship to des-
cribe size distribution. The method was not satisfactory for
this case and the visual estimation of slopes was used instead.
-------
SECTION 3
SOURCE AND CONTROL SYSTEM
The emission source for this test was a No. 7 gray iron
cupola used for an 8 to 9 hour/day operation. The cupola nor-
mally is run for 3 days/week, but during the test period it
was run for as many days as possible (3 days for the first
week, 5 days for the second week). The operation produces
10 to 12 tons per hour of metal from a feed of scrap metal,
coke, and limestone.
During the testing period the cupola ran at normal con-
ditions. The particles result from the melting of the scrap
and burning of the coke in the cupola. Figure 1 shows a sche-
matic of the entire operation including the scrubber system.
The hot gases from the melting operation enter the after-
burner where they are heated to approximately 700-1000°C. The
gases are then quenched by city water, which during the test
had a flow rate which varied between 0 and 2.6 £/s (4.1 GPM)
giving a temperature range of 700°C to 60°C. Scrubber liquor
is added just ahead of the throat section (the rate was varied
during the testing period from 2.5 H/s to 15.8 £/s) . Next the
gases enter a variable throat scrubber section which is a rect-
angular cross-sectioned orifice with a throat height of 91.4 cm
(36 inches). The throat width can be varied from 4.2 cm (1.7 in.)
to 11.4 cm (4.5 in.). Following the venturi the gases enter a
cyclone entrainment separator. An induced draft blower moves the
gas through the scrubbing system, powered by a 298 kW (400 HP)
motor. From this blower the gases are exhausted through an 11.9
m (39 ft) high, 0.91 m (3 ft) diameter stack.
At the venturi section, scrubber liquor is pumped in through
pipes placed directly upstream of the trim tabs. This liquor be-
-------
WET
CAP
CHARGING
DOOR
SCRUBBING
LIQUID
ENTRAINMENT
SEPARATOR
CUPOLA
AFTERBURNER QUENCHER
SCRUBBER
SUMP
FAN STACK
Figure 1. Schematic drawing of scrubber system,
-------
gins as clean water every day from a sump below the cyclone. No
fresh water feed is added to this liquor so that as the day pro-
gresses the solids concentration increases. When the system is
shut down the sump is neutralized, cleaned out, and fresh water
pumped in for the next day of operation.
The water used for the quench sprays, however, comes directly
from the city water lines and no recirculation is used.
-------
SECTION 4
TEST METHOD
The performance of this gas-atomized spray scrubber was
determined from simultaneous measurements of particle size
distributions and mass loading of the inlet and outlet gases.
Modified EPA-type sampling trains with heated University
of Washington Mark III (U.W.) cascade impactors were used
during this test period for particle measurement above 0.2
ymA dia. Figure 2 shows a schematic diagram of the modified
sampling train. Outlet impactor runs were performed in-stack,
while the inlet runs were ex-stack. The high gas temperatures
at the inlet sampling ports prevented in-stack sampling. Glass
fiber filter (Gelman type AE) substrates were used in the
impactors to prevent particle bounce and minimize wall losses.
On the inlet runs low velocity stages were used in the impactor
to increase the length of the sampling time.
The Air Pollution Technology portable screen diffusion
battery (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 (CNC) was used to determine
the particle number concentration at several locations in the
battery. From these data the size distribution may be deter-
mined for the particles smaller than 0.1 ym. The size distri-
bution computation was based on a calibration of the S.D.B.
performed in the A.P.T. laboratory.
-------
oo
PRECUTTER
AND
NOZZLE
u
THERMOMETER IMpINGER TRAIN'
HEATED
CASCADE
IMPACTOR
ORIFICE
METER
STA
7
*
c
b-
K
i
MANOMETER
WALL
I
THERMOMETERS
ITlROTAMETER
VACUUM
GAUGE
DRY GAS
METER
VACUUM
PUMP
SILICA
GEL
DRYER
Figure 2 - Modified EPA sampling train with in-stack cascade impactor
-------
During an impactor run, inlet or outlet fine particle
size measurements were taken with the portable diffusion
battery. Since this cupola has a charging operation to keep
the data fairly constant, all impactor runs were started at
the beginning of a charge. The diffusion battery data were
also taken at the beginning of these charges so that inlet
and outlet S.D.B. measurements at different times were con-
sidered to approximate simultaneous sampling.
Several impactor blank runs were performed on the inlet
and outlet simultaneously to ensure that the stack gases did
not react with the substrates. A blank impactor run consisted
of a heated impactor preceeded by two heated glass fiber filters
run at identical sample conditions as the other regular impactor
runs.
Gas flow rates for all tests were determined by means of a
calibrated S-type pitot tube. Velocity traverses of the inlet
and outlet were performed according to EPA Methods 1 and 2, and
average velocity points were selected for single point cascade
impactor sampling. Sampling flow rates were maintained with
the usual EPA train instruments (rotameter, dry gas meter, and
orifice manometer) to obtain isokinetic sampling throughout the
entire test. Orsat analyses were performed periodically on both
the inlet and outlet gases.
The inlet sample ports, because of a limited length of
ducting after quencher and before the venturi throat, were
located 0.73 equivalent duct diameters downstream of a 78° bend
and 0.73 equivalent duct diameters from the beginning of the ven-
turi section. The inlet duct rectangular (0.91 m x 0.61 m) and
its equivalent duct diameter was 0.73 m as calculated by EPA
Method 1.
The outlet sample ports were located in a 0.91 m (3 ft) dia-
meter round stack, about 8 stack diameters downstream of the inlet
from the fan and 3 diameters upstream of the stack exit. The vel-
ocity traverses indicated a well-developed flow pattern.
-------
SECTION 5
PROCESS CONDITIONS
During a two-week testing period 17 impactor sampling runs
(two of them being blank impactor runs} were made. The scrubber
process conditions for the inlet and the outlet are shown in
Tables 1 and 2. Runs 1 through 4 represent the scrubber running
with a number of leaks, as can be seen by comparing the inlet
and outlet flow rates. Between runs 4 and 5 many of the large
leaks were patched, resulting in much smoother operation as well
as a reduction of the fan amperage, an increase in the inlet flow
rate, and reduction in the outlet flow rate.
After patching most of the leaks, more control over the
scrubber operation could be maintained. It was then possible
to adjust water flow rates as well as the pressure drop across
the venturi. This enabled a number of test conditions to be ex-
amined. For the most part, the charging operation was fairly con-
stant and did not vary significantly between the runs, except for
run 1 when the crane broke and only one charge was added during
the entire run. The average barometric pressure during the testing
period was 99.36 kPa (29.34 in. Kg)
The inlet humidities from runs 2,9, and 10 were derived from
dessicant and impinger weight gains in the sampling train. The
inlet humidities for all other runs, except 12 and 14, were calcu-
lated as follows:
In runs 9 and 10, when the quencher water was off, the inlet
gas temperature was about 700°C (1,300°?), its humidity about 2%.
These conditions correspond to a wet bulb (saturation) temperature
of 68°C (155°F). The humidities at inlet temperatures above 68°C
were then calculated from a psychometric chart, assuming a wet
bulb temperature of 68°C. The humidity for runs 12 and 14 were
10
-------
TABLE 1. SCRUBBER CONDITIONS
Run
No.
1
2
3
4
5
6
1-B
7
8
2-B
9
10
11
12
13
14
15
P
i
-1.8
-1.3
-3.0
-2.0
-6.4
-6.4
-6.6
-6.1
-3.8
-3.6
-5.1
-4.3
-2.3
-2.5
-2.8
-5.8
-5.8
P
2
- 77.5
- 81.3
- 57.2
- 64.8
- 83.8
- 85.1
- 87.6
- 87.6
- 86.4
- 85.1
- 87.6
- 85.1
-105.0
-105.0
-107.0
- 94.0
- 92.7
P
3
*
- 83.8
- 94.0
- 63.5
- 78.7
- 96.5
- 94.0
- 96.5
- 94.0
-107.0
-109.0
-109.0
-107.0
-122.0
-125.0
-125.0
-109.0
-109.0
P
-137.0
-147.0
-137.0
-135.0
-114.0
-114.0
-117.0
-114.0
-119.0
-122.0
-124.0
-122.0
-132.0
-130.0
-132.0
-117.0
-117.0
P
5
0.9
0.9
1.1
1.1
-0.1
-0.1
-0.1
-0.1
-0.1
-0.1
-0.2
-0.2
-0.1
-0.1
-0.1
-0.1
-0.1
AP
75.7
80.0
54.2
62.8
76.4
78.7
81.0
81.5
82.6
81.5
82,5
80.8
103.0
104.0
104.0
88.2
86.9
Throat
Width
(cm)
11.4
4.4
5.6
4.4
11.4
11.4
11.4
11.4
5.1
5.1
5.6
5.6
4.2
4.3
4.3
11.4
11.4
Fan
Load
(kW)
310
276
285
294
245
242
245
242'
217
223
220
223
211
217
217
236
232
Scrubber
Liq.Flow
(A/s)
3.
11.
< 2.
< 2.
12.
12.
12.
12.
9.
9.
9.
9.
12.
12.
12.
15.
15.
2
4
5
5
6
6
6
6
5
5
5
5
6
6
6
8
8
Quench
Flow
U/s)
1.1
0.0
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
0.0
0.0
2.6
2.5
2.5
1.7
1.7
P = static pressure before venturi AP = P - P
i i :
P = static pressure after venturi
2
P = static pressure before fan dampers
3
P = static pressure after fan dampers
P = static pressure at outlet sampling ports
5
* = numbers expressed in cm W.C.
-------
TABLE 2. PROCESS CONDITIONS
Run
No.
1
2
3
4
5
6
1-B
7
8
2-B
9
10
11
12
13
14
15
Inlet
Temp.
°C
246
496
132
132
177
149
149
177
177
177
699
677
68
60
71
66
177
Outlet
Temp.
°C
56
56
57
57
65
65
65
54
54
54
58
58
51
51
51
51
51
Inlet Outlet
Water Water
Vapor Vapor
Vol.% Vol.*
**
21
11(6*)
27
27
25
26
26
25
25
25
2*
2*
27
20
32
25
25
7
7
17
17
20
20
20
5
5
5
6
6
8
8
8
11
11
Inlet
Flow Rate
Am3/s
(MACFM)
5.2(11.0)
6.2(13.2)
6.5(13.7)
6.5(13.7)
11.4(24.1)
11.0(23.4)
11.0(23.4)
10.8(22.9)
10.8(22.9)
10.8(22.9)
12.6(26.7)
12.5(26.4)
4.1( 8.7)
4.0( 8.4)
4.2( 8.8)
9.1(19.2)
10.5(22.2)
Outlet
Flow Rate
Am3/s
(MACFM)
***
11.7(24.8)
11.7(24.8)
12.5(26.5)
12.5(26.5)
10.5(22.2)
10.5(22.2)
10.5(22.2)
10.0(21.1)
10.0(21.1)
10.0(21.1)
8.1(17.1)
8.1(17.1)
7.0(14.8)
7.0(14.8)
7.0(14.8)
9.8(20.7)
9.8(20.7)
* Based on sampling train water catch
** Based on wet and dry bulb temperatures
*** MACFM = thousand actual cu.ft/min
12
-------
those based on the inlet being saturated.
The inlet sampling was done out-of-stack at a lower temp-
erature and condensation could have occurred in the pre-cutter,
probe, etc. Run 2 was the only run for which the calculated
humidity was greater than that which was based on evaporation
of the quench water. At the outlet, wet and dry bulb thermo-
meters were used to measure stack humidity.
The top of the cupola was open, except for a "wet cap".
The wet cap consists of baffles and sprays at the top of the
cupola and it was operated during the test period. Depending
upon the scrubber conditions, varying quantities of visible
emission were observed flowing from the wet cap.
The tests were conducted during the winter so that the am-
bient temperature averaged about 0°C during the testing. This
low ambient temperature probably caused the temperature of the
gases entering the scrubber to be lower than normal because of
heat transfer through duct walls, leakage of cold air, and use
of cooler than normal quencher and scrubber water.
The cyclone separator was not operating efficiently during
the testing period. It was patched along with the rest of the
system after run 4 but it could not be sealed completely as it
was in a rather rusted and patchwork condition. Any leakage
into the cyclone would disrupt the flow of liquid down the walls
and would cause reentrainment. Reentrainment caused by leakage
and perhaps poor flow distribution at the inlet, and internal
roughness due to scale build-up was clearly noticeable as rain-
out from the stack.
The Orsat analyses of the outlet gases are presented in
Table 3.
13
-------
TABLE 3. AVERAGE GAS COMPOSITION
Gas
Component
N2
02
C02
CO
Molecular Wt.
Inlet
Volume, %
Dry
80.1
6.4
13.2
0.3
29.9
Outlet
Volume, %
Dry
(Runs 1-4)
80.1
16.9
3.0
0
28.8
Outlet
Volume, %
Dry
(Runs 5-15)
79.6
14.6
5.8
0
29.2
14
-------
SECTION 6
CASCADE IMPACTOR PARTICLE DATA
Particle size distribution data were obtained for the gas-
atomized spray scrubber as described in the Test Method section.
Concurrent single point sampling at the average velocity loca-
tion was performed at the inlet and outlet. However, due to
the different mass loading rates the sample duration differed
between inlet and outlet. The average sampling times for the
inlet and outlet, respectively, were 6.5 and 31 minutes.
Entrained water drops were a problem at both the inlet and
outlet, so pre-cutters were used ahead of the impactors
during the sampling runs. The aerodynamic cut diameters of
these pre-cutters for the inlet and outlet were approximately
11.6 vmA and 5.6 ymA, respectively. Both the inlet and outlet
sampling were approximately isokinetic.
Isokinetic conditions are not crucial for sampling fine
particles. For example, the error caused by sampling 4 ymA
particles at a velocity 50% higher or lower than the gas stream
velocity would only be about 2 or 31 of the concentration.
To minimize the possibility of condensation in the outlet
impactors and to collect dry particles, the impactors (except
run 9 which was below the stack temperature due to a malfunc-
tion of the heating blanket) were maintained at about 13.5°C
above the gas stream temperature by heating blankets. Because
of the temperature fluctuation at the inlet the impactor was
kept out of the stack and heated by a heating blanket to an
average temperature of 86°C. The pre-cutters in both cases
(inlet and outlet) were not heated.
The fact that the impactors were heated should be noted
when interpreting the size distribution data. Because of the
15
-------
presence of a quencher upstream of the scrubber section, the
particles may be wet due to condensed or absorbed water when
they enter the scrubber. The wet particles would have a
different size distribution than those collected in a heated
impactor. The wet particle size distribution would have
fewer smaller diameters and probably a larger mass median diam-
eter. The runs with the most favorable conditions for wet par-
ticles were runs 11-14.
Particle concentration, particle size, and sampled volume
for cascade impactor runs are tabulated in the appendix in
Tables A-l through A- 15. Size distributions for the impactor
runs are given in Figures A-l through A-15 located in the
appendix. Table 4 summarizes the "d " and "a " results for
sr O o
the cascade impactor runs using log-probability parameters.
The distribution data include only the results of the cascade
impactor stage analysis. The pre-cutter and probe weight
gains were not used.
WE I BULL DISTRIBUTION
The Weibull distribution is an exponential function with
three parameters that make it more versatile than the log-
probability function. A description of the Weibull distribu-
tion function is given in Appendix "B". Table 5 summarizes
the Weibull distribution data and includes the calculated
mass median diamter (d ). The linear correlation coefficients
P s o
are all very good. These parameters were used to plot the
Weibull function on the figures in Appendix "A".
AERODYNAMIC PARTICLE DIAMETER
In this report the particle size determined from cascade
impactor analysis is presented in units of "ymA". This diam-
eter is the aerodynamic resistance diameter as described in
a discussion of aerodynamic diameter convention by Raabe
(1976). The relation to geometric diameter is
V • dP
16
-------
where d = aerodynamic resistance diameter, ymA
d = geometric diameter, ym
p = particle density, g/cm3
C1 - Cunningham slip correction factor
17
-------
TABLE 4. CASCADE IMPACTOR DATA USING
LOG-PROBABILITY ANALYSIS
Run No .
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Inlet
dPg
(pmA)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.47
.45
.40
.26
.19
.33
.39
.45
.58
.37
.40
.41
.31
.20
.18
Inlet
2
2
2
. 2
3
2
2
2
2
2
2
2
2
4
4
.6
.9
.1
.7
.4
.0
.0
.0
.0
.7
.4
.4
.7
.4
.4
Outlet
(ymA)
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
86
76
38
58
36
43
41
39
49
35
53
43
41
43
49
Outlet
1
4
2
2
2
2
2
2
2
2
2
2
2
2
2
.9
.0
.3
.3
.3
.2
.2
.5
.3
.7
.0
.7
.4
..4
.2
18
-------
TABLE 5. CASCADE IMPACTOR DATA
USING WEIBULL ANALYSIS *
Run dpso dpo e b LCC**
No. Type (ymA) (ymA) (ymA)
1 In 0.48 0.06 0.677 0.926 0.99
1 Out 0.72 0.39 1.12 0.457 1.0
2 In 0.45 0.07 0.673 0.806 0.98
2 Out 0.78 0 1.24 0.783 1.0
3 In 0.37 • 0.13 0.506 0.798 0.98
3 Out 0.36 0.27 0.450 0.510 1.0
4 In 0.26 0.11 0.368 0.682 0.99
4 Out 0.52 0.26 0.711 0.677 1.0
5 In 0.18 0.10 0.266 0.509 0.99
5 Out 0.34 0.29 0.416 0.427 0.99
6 In 0.30 0.14 0.405 0.767 0.98
6 Out 0.38 0.28 0.484 0.487 1.0
7 In 0.33 0.17 0.445 0.657 0.99
7 Out 0.36 0.27 0.460 0.459 1.0
8 In 0.46 0.12 0.612 0.967 0.99
8 Out 0.36 0.33 0.422 0.322 1.0
9 In 0.52 0.28 0.704 0.630 0.99
9 Out 0.43 0-29 0.585 0.501 0.99
10 In 0.32 0.23 0.444 0.451 0.99
10 Out 0.34 0.27 0.432 0.408 1.0
11 In 0.38 0.12 0.542 0.734 0.98
11 Out 0.45 0.33 0.594 0.487 0.99
12 In 0.36 0.16 0-518 0.643 0.98
12 Out 0.4.0 0.26 0.545 0.519 1.0
13 In 0.29 0.12 0.424 0.648 0.99
13 Out 0.37 0.28 0.488 0.457 1.0
14 In 0.19 0,06 0.325 0.512 0.99
14 Out 0.38 0.28 0.506 0.473 1.0
15 In 0.17 0.09 0.282 0.438 0.98
15 Out 0.42 0.29 0.571 0.471 0.99
* Column headings are defined in Appendix B
** Linear correlation coefficient
19
-------
SECTION 7 •
DIFFUSION BATTERY DATA
Diffusion battery data were taken during the second week
of testing. Inlet data were taken during runs 5, 6, 1-B, 8,
2-B, 12, and 13. Outlet data were taken between runs 2-B and
9 and during run 11.
A dilution system was used for this testing along with
the A.P.T. diffusion battery. A 208 liter (55 gallon) drum
was used upstream of the diffusion battery as this dilution
system. A sample should not be taken continuously from the
stack because the intermittent charging operation can result
in rapidly changing readings. A syringe was inserted into the
stack several times to withdraw gas samples and these were ex-
pelled into the dilution tank. Later this system was switched
to a calibrated hand pump. These gas samples were diluted with
dry filtered air in the tank and run through the diffusion bat-
tery with particle counts being measured with the condensation
nuclei counter.
The particle size distributions determined by diffusion
battery analysis are given in Table 6. The diffusion battery
results are averages of a number of counts during each run and
are reduced using a log-normal distribution fit to penetration
through successive screens. The geometric count or number mean
diameter (d N) and geometric standard deviation are the log-
normal parameters used in Table 6. The cut sizes of the screens
ranges between 0.03 ym and 0.23 ym.
20
-------
TABLE 6. DIFFUSION BATTERY PARTICLE
SIZE DISTRIBUTIONS
D.B.
Run
No.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Impactor
Run No.
5
5
6
1-B
8
8
2-B
2-B*
2-B*
2-B*
11
11
12
13
Type
of
Run
In
In
In
In
In
In
In
Out
Out
Out
Out
Out
In
In
dpN
(ym)
0.070
0.092
0.095
0.128
0.072
0.085
0.094
0.057
0.061
0.085
0.090
0.090
0.086
0.059
CTg
3.4
12.0
12.0
14.0
6.0
2.9
11.5
1.0
2.5
4.5
6.0
8.9
9.3
6.0
Total
Particle
Count **
No. /cm3
9.0 x 106
1.1 x 107
1.1 x 107
3.9 x 106
2.3 x 10s
1.3 x 106
1.1 x 107
5.1 x 106
8.0 x 106
8.3 x 10s
7.8 x 10s
1.9 x 106
2.2 x 107
1. 5 x 107
* Same condition but not simultaneous.
** Gas conditions: 20°C, 1 atm
21
-------
SECTION 8
PARTICLE PENETRATIONS
Particle penetrations versus particle aerodynamic
diameter were calculated from cascade impactor and diffusion
battery cumulative loading data. The penetrations based on
diffusion battery data were calculated from log-normal inlet
and outlet cumulative distributions. The physical (actual)
size distributions from the diffusion battery analysis were
converted to aerodynamic by assuming the cupola dust had a
particle density of 2.5 g/cm3.
Calculation of the particle penetration from cascade
impactor data was tried using the mathematical relations of
the Weibull distribution function. This method was aban-
doned because in most cases the calculation gave the result
that penetration increased as particle size increased. Also,
there was often a minimum in the size range between 1 and
2 ymA diameter. The unrealistic penetration curves resulted
inspite of the relatively high linear correlation coefficients
of the size distributions.
Since the size distributions were also not log-normal, the
cascade impactor penetrations presented in Figures 3-9 were calcu-
lated graphically. The method involved graphically determining
the slope of the cumulative mass loading versus aerodynamic
particle diameter curve, drawn from the data presented in
Appendix "A". The ratio of the slopes of the outlet curve to
the inlet curve at a certain particle size was the penetra-
tion for that particle size.
The total mass loadings and overall penetration for the
runs are presented in Table 7 . Mass loading and penetration
22
-------
are also shown without the outlet pre-cutter catch. The reason
being that the pre-cutter loading was mostly entrainment carry-
over from the cyclone and was not due to scrubber inefficiency.
23
-------
TABLE 7. MASS LOADING AND
OVERALL PENETRATION
Run
No.
1
2
3
4
5
6
1-B
7
8
2-B
9
10
11
L2
L3
L4
L5
Mass
Inlet
1,570
2,120
1,100
2,060
3,190
2,340
7,030
2,530
1,913
1,600
5,070
2,620
1,750
2,370
2,500
1,450 •
1,240
Loading,
Outlet1
212
74.3
280
176
609
341
968
234
320
177
198
589
159
214
797
278
277
mg/DNm3 *
Outlet2
46.3
36.5
267
175
514
307
-
234
98.2
-
156
159
153
159
174
186
223
Penetration, 1
Pt1
13.5
3.5
25.5
8.5
19.1
14.6
13.8
9.3
16.7
11.1
3.9
22.5
9.1
9.0
31.8
19.2
22.2
Pt2
2.9
1.7
24.4
8.5
16.1
13.1
-
9.2
5.1
-
3.1
6.1
8.7
6.7
6.9
12.8
17.9
* Including outlet pre-cutter dry weight gain
Without outlet pre-cutter dry weight gain
* N = 0°C, 1 atm
24
-------
i.o
.01
0.2
0.5
1.0
2.0
5.0
10.0
PARTICLE AERODYNAMIC DIAMETER, d , ymA
pa
Figure 3. Particle penetration for cascade impactor
runs 1 and 2.
25
-------
1.0
0.5
u
0.2
g 0.1
w
CL,
n c
.05
,02
01
0.2
0.5 1.0
2.0
5.0 10.0
PARTICLE AERODYNAMIC DIAMETER, d , ymA
pa
Figure 4. Particle penetration for cascade impactor
runs 3 and 4.
26
-------
01
.05
0.1
10.0
PARTICLE AERODYNAMIC DIAMETER, d , ymA
pa
Figure 5. Particle penetration for cascade impactor runs
5, 6 and 7, and diffusion battery runs 3-9 (inlet)
and 10-12 (outlet).
27
-------
.05
PARTICLE AERODYNAMIC DIAMETER, d , ymA
pa
Figure 6. Particle penetration for cascade impactor run 8
28
-------
1.0
e
2
UH
o
H
pq
W
10.0
PARTICLE AERODYNAMIC DIAMETER, d , umA
pa'
Figure 7. Particle penetration for cascade impactor
runs 9 and 10.
29
-------
2
O
t—i
H
U
2
O
H
W
,005
5,0 10.0
Figure 8
PARTICLE AERODYNAMIC DIAMETER, d umA
pa'
Particle penetration for cascade impactor runs
11, 12 and 13, and diffusion battery runs 15
and 16 (inlet) and 13 and 14 (outlet).
30
-------
1.0
0.5
o
I—I
I
UH
«\
o
I-H
H
ta
z
,02
01
0.2
0.1 -H": r
. 0 5 - -IT
5.0
10
0.2 0.5 1.0 2.0
PARTICLE AERODYNAMIC DIAMETER, d^Q,
pa
Figure 9. Particle penetration for cascade impactor runs
14 and 15.
31
-------
SECTION 9
OPERATING PROBLEMS
The primary operating problem with the system at first
was the limiting capacity of the fan. Great amounts of ex-
cess air were getting into the system through all of the
leaks, resulting in the fan running at its maximum capacity.
Because of this factor the system could not pull enough gas
through the scrubber and a large portion exited through the
top of the cupola. The amount of water that could be added
to the system was also limited which in turn affected the
scrubber efficiency.
Once most of these leaks were patched, however, the fan
amperage dropped well below the limit, and the system could
be adjusted easily.
The scrubber liquor was another operational problem.
Since the cyclone entrainment separator was not working ef-
fectively, much of the water used to capture the particles
was carried over and through the fan and out the stack. The
scrubber liquor was recycled and changed only at the end of
each day, which meant that the solids concentration increased
as the day progressed. Reentrainment of this water, therefore,
would mean higher outlet loadings since this extra weight was
considered in the emission rates.
Corrosion had a severe effect on the system operation
since it was the cause of the excessive leakage of air into
the scrubber and entrainment separator.
Another problem with the system was the absence of a cap
on the cupola top. This allowed the gases to escape through
the top. Not all the gases pass through the scrubber and
32
-------
therefore are not treated effectively, which in turn caused a
visible emission problem. This can be alleviated by capping
the cupola top.
33
-------
SECTION 10
ECONOMICS
The scrubber system was built about ten years ago with
substantial participation by the foundry. Consequently, the
cost was distributed among several purchased components, con-
tracted work, and foundry performed work. Records of the costs
were not kept.
The operating cost for the blower (300 kW motor) was about
$72.00 per day, based on eight hours use per day.
34
-------
SECTION 11
PERFORMANCE COMPARISON
The performance of venturi-type scrubbers has been modeled
extensively. A recent survey and the model used here are pre-
sented in Yung, et al. (1977) which is summarized in Appendix "C".
The scrubber parameters are known or estimated from the
average of the conditions for the sets of runs shown in Table 8.
The velocity attained by the liquid drops at the throat exit is
often as high as 801 of the gas velocity. However, the pressure
drop predicted based on this velocity accommodation is much high-
er than measured. The model predicts that the throat length for
this 80% velocity attainment by drops should have been about 0.8
m. Since the actual throat length was only about 0.1 m the lower
measured pressure drop is reasonable. For a throat length of 0.1
m the drops attain only about 60% of the gas velocity for most of
the runs according to the model. And the predicted pressure drops
more closely match those measured.
The predicted particle penetrations are shown compared to the
average of the run sets in Figures 10-14. The solid (measure-
ment) curves represent the measured penetration without account-
ing for leakage. The dashed curves assume that the outlet was
diluted by the factor listed in Table 8 which is based on the
data in Table 2. Table 9 presents the predicted pressure drop
for a 0.1 m throat length and the overall penetration for d =
Jr o
0.4 ymA, a =2.5. These values may be compared to the measured
c^
values, presented in Tables 7 and 8.
The comparison of predicted and measured penetration leads
to the following observations:
1. The predictions are generally close to the experimental
results for particles around 1 jamA diameter. Runs 5, 6, and 7
35
-------
TABLE 8. AVERAGE CONDITIONS FOR PERFORMANCE PREDICTIONS
Run Set
5
9
11
14
,6,7
8
,10
,12,13
,15
AP
cm W.C.
79
83
82
103
88
Q
m3
11
10
12
4
9
G
/s
.1
.8
.5
.1
.8
m3/m3
0
0
0
0
0
.0011
.0009
.0008
.0031
.0016
ut
m/s
106
232
244
104
104
T
LG
°C
168
177
688
66
122
PG
kg/m3
0.
0.
0.
0.
0.
72
70
36
93
80
kg/m-s
2.
2.
4.
1.
2.
IxlO"5
2xlO"5
IxlO"5
7xlO"5
OxlO"5
Dil*
1
1
1
1
1
.23
.27
.88
.79
.22
*"Dil" is the ratio of measured outlet to inlet volume flow rates at standard conditions.
-------
TABLE 9. PREDICTED PRESSURE DROP
AND OVERALL PENETRATION
Run Set
5,6,7
8
9,10
11,12,13
14,15
AP
cm W.C.
78
349
342
148
91
Penetration
%
47
26
33
44
45
Assumptions: Throat Length = 0.1 m
d =0.4 ymA
Pg
a = 2.5
g
37
-------
1.0
S5
o
o
I—I
H
W
0.1
With Dilution Factor
0.05
0.02
0.01
0.1
0.2
PARTICLE AERODYNAMIC DIAMETER, d , ymA
pa'
10.0
Figure 1Q. Comparison of predicted with measured
penetration for average of Runs 5, 6,
and 7.
38
-------
25
o
I-H
H
U
o
I—I
H
W
With Dilution Factor
5.0
PARTICLE AERODYNAMIC DIAMETER, d,umA
pa
10.0
Figure li. Comparison of predicted with measured
penetration for Run 8.
39
-------
0.01
0.1
0.2
0.5
1.0
2.0
5.0
10.0
PARTICLE AERODYNAMIC DIAMETER, d ,umA
pa'
Figure 12. Comparison of predicted with measured
penetration for average of Runs 9 and
10.
40
-------
§
I—I
H
U
o
I—I
H
W
W
With Dilution Factor
10.0
PARTICLE AERODYNAMIC DIAMETER, dpa>ymA
Figure 13. Comparison of predicted with measured
penetration for average of Runs 11, 12,
and 13.
41
-------
2
O
H
M
2
W
With Dilution Factors!
o.os ia
0.02
0.01
10.0
PARTICLE AERODYNAMIC DIAMETER, d , umA
pa'
Figure 14. Comparison of predicted with measured
penetration for average of Runs 14
and 15.
42
-------
have the poorest fit. Although the operating conditions for
these runs were similar to those of runs 14 and 15, the exper-
imental penetration was greater. This may be due to the fact
that the gas flow rate was higher in runs 5, 6, and 7 which
could mean more leakage and could cause more entrainment carry-
over at the outlet. The effect of L/G ratio on the model pre-
dictions is shown by comparing the predicted penetrations of
these two run sets. The main difference in the operating con-
ditions was that runs 14 and 15 had a 45% higher L/G ratio than
runs 5, 6, and 7.
2. Predicted penetration for sub-micron particles is
generally higher than measured. This may be due to condensation
caused by the cool duct walls and the cool scrubber liquid which
had low temperatures due to the cold ambient air (0°C). Leaks
of the cold ambient air into the system would also have caused
condensation.
3. Predicted penetration is generally lower than measured
for particles larger than about 1 ymA diameter. There are several
plausible causes of this disparity:
a. Entrainment not collected in the entrainment sepa-
rator would yield particles in the larger size range. The
cyclone separator was known to be inefficient although its
^performance was improved by internal modifications and leak
patching. This inefficiency was clearly demonstrated by
noticeable liquid drops at the outlet stack where the
sampling crew was working, and by the pre-cutter mass
collections.
b. Imprecise knowledge of the liquid flow rate can
cause the prediction to be substantially incorrect in the
larger particle size range. This effect becomes more
serious as the liquid/gas ratio decreases, as seen by
comparing the prediction of runs 5, 6, and 7 with that
of runs 14 and 15.
4. Predicted overall penetration was generally higher than
measured. This difference is probably due primarily to the as-
sumption that the size distribution is log-normal with the same
standard deviation for all particles less than 0.4 ymA (d ).
43
-------
Condensation probably affected the log-normality of the small
size distribution. Also, the penetration model over-predicted
the penetration of the smaller particles which would definitely
cause the overall penetration prediction to be too high.
In conclusion, with all the real effects considered, the
model provides an adequate prediction. Predictions for several
of the runs, notably 8, 11, 12, 13, 14, and 15, were quite good.
44
-------
REFERENCES
Lipson, C., and N.J. Sheth, "Statistical Design and Analysis
of Engineering Experiments," McGraw-Hill, 1973.
Raabe, O.G., "Aerosol Aerodynamic Size Conventions for
Inertial Sampler Calibration," J. Air Poll. Control
Assoc., 26^:856-860, 1976.
Yung, S., S. Calvert, and H.F. Barbarika, "Venturi Scrubber
Performance Model," A.P.T., Inc., San Diego, California.
EPA Contract No. 68-02-1328, Task No. 13, July 1977.
45
-------
APPENDIX A
Cascade Impactor Data
46
-------
TABLE A-l. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #1
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
1,570
1,360.
1,360
1,360
1,320
1,260
862
556
375
d
pc
(ymA)
12.40
46.80
3.98
2.28
1.27
0.76
0.42
0.23
0.070
OUTLET
M
cum
(mg/DNm3)
212.0
46.3
43.6
41.3
41.3
37.4
30.5
20.8
6.2
d
PC
(ymA)
6.57
25.0
11.0
5.19
2.12
1.22
0.67
0.40
0.260
M
cum
cumulative mass collected on that stage and those
below
cut diameter (aerodynamic) for that stage
2
ymA = microns, aerodynamic = d_(Cf p )V
d = particle diameter (actual)
C1 = Cunningham correction factor
p = particle density (g/cm3)
DNm3 = dry normal cubic meters (0°C, 1 atm)
47
-------
TABLE A-2. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #2
IMPACTOR
STAGE
NUMBER
Precutter
<| Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
2,120
1,800
1,770
1,750
1,710
1,630
1,090
530
d
PC
(vmA)
12.1
45.4
3.86
2.21
1.23
0.71
0.40
0.22
0.030
OUTLET
M
cum
(mg/DNm3)
74.3
36.5
35.9
35.9
34.2
27.4
22.3
17.5
11.9
d
PC
(pmA)
6.35
24.90
10.90
4.21
2.11
1.21
0.67
0.38
0.520
oo
TABLE A-3. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #3
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
Mcum
(mg/DNm3)
1,100
833
833
833
824
790
652
357
219
(umA)
11.70
43.80
3.73
2.14
1.19
0.71
0.39
0.21
0.020
OUTLET
Mcum
(mg/DNm3)
280
267
267
266
264
256
238
187
116
(umA)
5.41
21.20
9.30
3.97
1.80
1.03
0.57
0.33
0.380
TABLE A-4. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #4
IMPACTOR
STAGE
NUMBER
Precutter
$ Nozzle
1
2
• 3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNma)
2,060
1,710
1,710
1,700
1,680
1,630
1,460
936
691
d
PC
(vmA)
11.40
42.80
3.64
2.09
1.16
0.67
0.38
0.20
0.020
OUTLET
M
cum
(mg/DNm3)
176
175
174
171
162
143
91.7
47.4
d
V
(umA)
S.84
22.30
9.76
3.77
1.89
1.08
0.60
0.34
0.610
TABLE A-5. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #5
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
Mcum
(mg/DNm3)
3,190
1,570
1,570
1,550
1,520
1,470
1,350
1,000
800
V
(umA)
10.10
38.00
3.23
1.85
1.03
0.62
0.33
0.18
0.020
OUTLET
"cum
(mg/DNm3)
609
514
513
511
506
490
466
365
238
d
pc
(vmA)
5.47
21.0
9.20
4.36
1.78
1.02
0.56
0.33
0.200
-------
TABLE A-6. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #6
IMPACTOR
STAGE
NUMBER
Pre cutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
Mcum
(mg/DNm3)
2,340
1,250
1,250
1,250
1,230
1,190
1,050
577
269
V
(umA)
10.10
38.10
3.24
1.86
1.03
0.59
0.34
0.18
0.010
OUTLET
Mcum
(mg/DNm3)
341
307
306
305
299
287
268
205
112
V
(umA)
5.47
21.10
9.24
3.57
1.79
1.02
0.57
0.32
0.300
TABLE A-7. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #7
IMPACTOR
STAGE
NUMBER
Precutter
6 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
"cum
(mg/DNm3)
2,530
1,460
1,450
1,430
1,410
1,360
1,190
627
246
V
(ymA)
10.50
39.60
3.37
1.93
1.07
0.62
0.35
0.19
0.010
OUTLET
Mcum
(mg/DNm3)
234
234
233
232
227
218
202
156
81.8
Sc
(ymA)
5.21
20.20
8.85
3.42
1.72
0.98
0.54
0.30
0.350
TABLE A-8. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #8
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
S
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
1,910
1,850
1,850
1,840
1,810
1,720
1,260
559
236
-------
TABLE A-10. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #10
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
2,620
1,740
1,710
1,680
1,660
1,590
1,460
1,050
652
V
(umA)
14.40
54.20
4.61
2.64
1.47
0.85
0.48
0.27
0.030
OUTLET
"cum
(mg/DNm3)
589
159
157
156
153
146
135
112
68.8
V
(ymA)
5.31
20.50
8.96
3.46
1.74
0.99
0.55
0.31
0.440
TABLE A-ll. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #11
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
1,750
1,270
1,270
1,250
1,230
1,180
942
563
388
d
PC
(UmA)
12.0
45.0
3.82
2.19
1.22
0.73
0.40
0.22
0.020
OUTLET
M
cum
(mg/DNm3)
159
153
152
151
148
143
132
87.5
45.5
d
PC
(umA)
5.64
22,40
9.82
4.65
1.90
1.09
0.60
0.36
0.570
TABLE A-1Z. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #12
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
2,370
1,610
1,600
1,570
1,540
1,450
1,270
711
404
d
pc
(umA)
11.70
43.90
3.73
2.14
1.19
0.68
0.39
0.21
0.030
OUTLET
M
cum
(mg/DNm3)
214
159
158
158
155
145
132
103
61.0
d
pc
(pmA)
5.68
21.60
9.48
3.66
1.84
1.05
0.58
0.33
0.620
TABLE A-13. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #13
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
Mcum
(mg/DNm3)
2,500
1,920
1,920
1,900
1,860
1,790
1,500
988
680
(wmA)
11.0
41.20
3.51
2.01
1.12
0.67
0.36
0.20
0.030
OUTLET
Mcum
(mg/DNm3)
797
174
174
173
170
161
149
112
71.5
V
(vmA)
5.42
21.0
9.21
4.36
1.78
1.02
0.56
0.33
0.570
-------
TABLE A-14. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #14
IMPACTOR
STAGE
NUMBER
Precutter
d Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNrn3)
1,450
1,300
1,290
1,260
1,200
1,130
1,056
760
531
d
pc
(ymA)
10.20
38.40
3.27
1.87
1.04
0.60
0.34
0.18
0.010
OUTLET
Mcum
(mg/DNm3)
278
186
185
183
180
171
156
118
59.8
V
(ymA)
5.43
20.40
8.92
3.44
1.73
0.99
0.54
0.31
0.470
TABLE A-15. 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)
1,240
1,240
' 1,210
1,200
, 1,180
1,140
1,040
794
713
d
PC
(ymA)
11.90
44.80
3.81
2.18
1.21
0.73
0.40
0.21
0.010
OUTLET
M
cum
(mg/DNm3)
277
223
221
218
213
202
184
127
65.0
d
v
(ymA)
5.33
20.20
8.84
4.19
1.71
0.98
0.54
0.32
0.330
51
-------
10
en
t-o
|
a
o.
1 °-5
(U
0.2
0.1
O INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
2 5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-l. Inlet and outlet size distribution for run #1.
O INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
0.1
10
20 30 40 50 60 70 80 90
MASS PERCENT UNDERSIZE
95
98 99
Figure A-2. Inlet and outlet size distribution for run #2.
-------
10
n)
o.
1 0.5
at
ia
0.2
0.1
OINLET
A OUTLET
LOG-NORMAL
--- WEIBULL
2 5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-3. Inlet and outlet size distribution for run *3.
10
CO
O,
Oi
W
H
I
o
0.5
0.2
0.1
INLET
OUTLET
LOG-NORMAL
--- WEIBULL
5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-4. Inlet and outlet size distribution for run #4.
-------
01
O INLET
OUTLET
LOG-NORMAL
WEIBULL
0.1
2 5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-S. Inlet and outlet size distribution for run #5.
10
a
P.
O
O
0.2
0.1
O INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
2 5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-6. Inlet and outlet size distribution for run »6.
-------
10
Cn
ca
Q.
•a
H
IH
1
0.5
0.2
0.1
OINLET
A OUTLET
LOG-NORMAL
--- WEIBULL
10
2 5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-3. Inlet and outlet size distribution for run #3.
en
(H
H
g
0.2
0.1
INLET
OUTLET
LOG-NORMAL
--- WEIBULL
2 5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-4. Inlet and outlet size distribution for run #4.
-------
en
O INLET
OUTLET
LOG-NORMAL
--- WEIBULL
0.1
2 5 10 20 30 40 SO 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-5. Inlet and outlet size distribution for run #5.
10
a
o.
B:
to
Q
U
§ 0.5
a
0.2
0.1
O INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
5 10 20 30 40 SO 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-6. Inlet and outlet size distribution for run #6.
-------
Cn
tn
O INLET
A OUTLET
LOG-NORMAL
10 C?l
0.1
10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
(J INLET
A OUTLET
LOG-NORMAL
Figure A-7. Inlet and outlet size distribution for run #7.
0.1
5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-8. Inlet and outlet size distribution for run #8.
-------
10
Ol
a
as
\a
H
(14
S 1
<
1
o
w 0.5
0.2
0.1
INLET
A OUTLET
LOG-NORMAL
5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
0.1
10
20 30 40 50 60 70 80 90
MASS PERCENT UNDERSIZE
95
Figure A-9. Inlet and outlet size distribution for run #9.
Figure A-10. Inlet and outlet size distribution for run #10.
-------
O INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
0.1
2 5 10 20 30 40 SO 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-ll. Inlet and outlet size distribution for run #11.
|
a
p.
at,
sa
H
o
o
oi
O INLET
A OUTLET
_ LOG-NORMAL
-- WEIBULL
0.5
0.2
0.1
^^^^^^J^J^^JJ^^^^^JJq^^^^^J^^g^ggg^ggggg^gJggg^gggg^^f^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^y^^^^^^^^^
2 5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-12. Inlet and outlet size distribution for run #12.
-------
cx.
tu
H
Q
O
t-H
s
<
on >•
00 §
a:
0.5
0.2
0.1
INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
5 10 20 30 40 SO 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
INLET
A OUTLET
LOG-NORMAL
--- WEIBULL
0.1
10
20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
95 98 99
Figure A-13. Inlet and outlet size distribution for run #13.
Figure A-14. Inlet and outlet size distribution for run #14,
-------
w
H
Q
u
i—i
s
o
0!
O INLET
A OUTLET
LOG-NORMAL
WEIBULL
0.2
0.1
5 10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure A-15. Inlet and outlet size distribution for run #15
59
-------
APPENDIX B
Weibull Distribution
60
-------
WEIBULL DISTRIBUTION
The Weibull distribution, Lipson (1973), offers two 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.
Cumulative Distribution Function
CDF = 1 - exp
-
b 1
-
(B-l)
where, d = particle diameter
minimum particle diameter
characteristic diameter
b = Weibull slope
po
0
The CDF has the property that:
CDF (0 = d ) = 0.632
P
The median particle diameter occurs when the CDF = 0.5, so
pg
1/b
Linear Transformation
Transformation to a linear form;
y = A + B x
(B-3)
61
-------
requires that:
y = In In
-CDF
CB-4)
- ln
- V
(B-5)
which means that:
A = -b In (0 - dpQ)
B = b
(B-6)
(B-7)
or,
0 = d + exp
po ^
b = B
A
B
(B-8)
(B-9)
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 regres
sion is performed oh the size distribution data. Note that,
0 < d_ < smallest diameter found in the distribution
— po
Density Function
The Weibull density function is the derivative of the CDF:
fd -d
b-1
£(d
e-d
exp
po
P_P_
e-d
po
(B-10)
62
-------
Penetration
The penetration is the ratio of the cumulative mass loading
distribution derivatives,
pt co - °u
CT,in fin Cdp>
where, CT = total mass concentration or loading
"out" - refers to outlet particle size distribution
"in" - refers to inlet particle size distribution
and, f(dp) - is defined by equation (B-10J
Minimum Particle Diameter, Physical Interpretation
"d " is the smallest diameter of the total distribution.
A value other than zero means that the data indicate that there
is a minimum particle size. This is physically reasonable be-
cause of the particle formation mechanisms and possible agglo-
meration 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
"G" by equation (B-2) .
Weibull 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.
63
-------
APPENDIX C
Venturi Performance Model
64
-------
VENTURI SCRUBBER PERFORMANCE MODEL
Yung, et al., 1977* 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
-J * UG cVv
where Pt(d ) = penetration for particles with diameter d ,
fraction
u = relative velocity between dust and drop,
cm/sec
Up = 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 C-l to
predict particle collection in a venturi was developed.
(3) Particle collection predicted by equation C-l
agrees satisfactorily with performance data.
(4) Most of the particle collection occurs in
the venturi throat. The solution to equation C-l for the
venturi throat, using the inertial collection efficiency
correlation, and assuming a zero initial drop velocity, is
*Yune, S., S. Calvert, and H.F. Barbarika, "Venturi Scrubber
Performance Model, EPA 600/2-77-172, A.P.T., Inc., San Diego,
California, 1977.
65
-------
In Pt (d^ , . . . / v. =
^ P'- _ ,. * . ^ A If 1-,,* 1-5 + A 7 1_u* °'5
B
1 ..# \ + j^ 7* I
|XLljl~ U./ 1
p o \ d / I
- 5.02
™h
- 5-02
Kpo _.. u r- J-_ v -po
(C-2)
f >
\ /T + 8\2 /T + 8\I/T + 8\2 1 0-Sf
1 4. 1JT /ijuX./Ij^OM/ij^Ol-il V.
where u§ = 2 jl -f—g—-j +(—g—Jl—g—J-1J f
Pt (d ) = penetration for particles with diameter d ,
fraction
B
ill r i /"t * i
"Do
QT = volumetric liquid flow rate, cm3/sec
p = liquid density, g/cm3
Li
PG = gas density, g/cm3
CD = drag coefficient obtained from the "standard curve
u5 = dimensionless drop velocity
= ^Gt
u, = drop velocity, cm/sec
uGt = gas velocity in the throat, cm/sec
K = inertial parameter based on throat velocity
66
-------
C1 » Cunningham slip factor
d = particle diameter, ym
p = particle density, g/cm3
VLQ = gas viscosity, poise
d, = drop diameter, cm
L = dimensionless throat length
*t So PG
2 dd PL
Equation C-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 C-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 Fa u (C-3)
where AP = pressure, cm W.C.
UG = gas velocity in the throat, cm/sec
QT = liquid flow rate, cm3/sec
Jji
Qr = gas -flow rate, cm3/sec
FX = correction factor, dimensionless
F4 - 5^- 2
UGt
1 - X'+ (*
3 *tCDo PG
16 dd PL
(X*- X)
20'5
67
-------
Uj = drop velocity at the exit of the throat, cm/sec
I = throat length or distance between liquid
injection point and the exit of throat, cm
d, = drop diameter, cm
PG = gas density, g/cm3
PT = liquid density, g/cm3
Li
C-r. - 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,
68
-------
TECHNICAL REPORT DATA
(flcase read Instructions on the reverse before completing)
EPA- 600/2 -77-209a
(.TITLE AND SUBTITLE
Gas-atomized Spray Scrubber Evaluation
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
October 1977
6. PERFORMING ORGANIZATION CODE
Seymour Calvert, Harry F. Barbarika, and
Gary M. Monahan
8. PERFORMING ORGANIZATION REPORT NO.
JD ADDRESS
Air Pollution Technology, Inc.
4901 Morena Boulevard (Suite 402)
San Diego, California 92117
10. PROGRAM ELEMENT NO.
1AB012; 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
Task Final; 2-8/77
14. SPONSORING AGENCY CODE
EPA/600/13
,6.SUPPLEMENTARY NOTES ffiRL-RTP project officer for this report is Dale L. Harmon, Mail
Drop 61, 919/541-2925. EPA-600/2-76-282 is an earlier report in this series.
16. ABSTRACT
The report gives results of fine particle collection efficiency measurements
of a gas-atomized spray scrubber, cleaning effluent gas from a No. 7 gray iron
cupola. Tests were made at several levels of pressure drop and liquid/gas ratio.
Particle size measurements on inlet and outlet gas streams were made with cascade
impactors and an A. P.T. screen-type diffusion battery. The particle mass collection
efficiency at a pressure drop of about 100 cm W.C. was 91% for particles with a mass
median diameter of about 0.4 micrometers A. Scrubber inlet gas flow rate varied,
because the cupola was operated with an open top, from 4.0 to 12.6 A cu m/s (8,500
to 27,000 acfm). Air leakage into the scrubber system caused serious operating
problems until most of the leaks were sealed. The penetrations for 1 micrometer A
diameter particles were about as predicted by the mathematical model; however,
smaller particle penetrations were lower, and larger particle penetrations were
higher, than predicted. The latter two effects are believed due to water condensation
effects and entrainment, respectively.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
T3B 13H~
11F
11G
14B
07A_
Air Pollution Atomizing
tron and Steel Industry Spraying
Dust
Measurements
3ray Iron
Furnace Cupolas
Grubbers
Air Pollution Control
Stationary Sources
Particulate
Collection Efficiency
Gas-atomized Spray
Scrubber
i. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
77
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
A Form 2220-1 (9-73)
69
------- |